New PowerMacs and 20" LCD

[pyrotoaster]

Originally posted by smackdesign
So the powermacs are now at 1.4 ghzs. If apple updates the line every 6 months we should see them updated next in July. I hope they release the 970 at macworld NY and not another speed bumbed moto chip. I'm crossing my fingers.
Two things:
1. Unless IBM puts the 970 in mass production ahead of schedule, they won't be in Powermacs by July. Apple will need computer ready to, or almost ready to, ship when they announce any departure from Motorola and the G4.
2. To my knowledge, Apple won't be participating in MWNY. If you go to the MacWorld Expo website (http://www.macworldexpo.com/), you'll notice that there's no planned keynote speech (which is something that SJ would do).

Personally, I think both the new Powermacs and the new LCDs are great. If there is another PM revision in six months, it'll be even more minor than this. I think we'll see a Powermac 970 announcement in the Fall (although I'd love to see 970 machines sooner).

[iJon]

Originally posted by hitman


well, you could always go for this thing instead...

http://www.m-audio.net/products/consumer/revolution_page1.php
its not really the 5.1 or 7.1 that gets me. i have a 5.1 set up on my pc with an audigy drive, personally i couldnt tell the difference with 5.1 surround sound and just stero sound coming out of each of the 5 speakers. its more the eax sound tech that really gets me. that eax makes a difference on my pc, if i hit someone in gta3 with that on with my right tire, i can hear in my right speaker their bones crushing as the cops come after me behind, you gotta check it out, really cool, plus the remote control you get free with the audigy is pimp, i love it.

iJon

[iJon]

Originally posted by digitalrampage
Single processor for $1499?

For the extra $200, the dual 867 was a much faster machine...

To add to this, NO superdrive in the Mid Range....

Bluetooth is an Optional adaptor...

I SAY, WELCOME TO the NEW SCREENS AND PRICING...

And someone needs to be smacked in the head at apple for the new powermac lineup...

FOOLS!

I sell this stuff, powerbooks, GREAT NEWS, Powermacs BIG LOSS
yes your right, well it depends how you look at it. the previous line up was very nice, but took away sales from the middle and high end. lot of my customers couldnt justify the money for an extra hard drive space and the small system bus upgrade. this way when they look at the lineup they think well for this much more i can get the middle instead of the low end. get what im saying. maybe bad for the guy buying if he is short on cash and wants a big bang for his buck but a smart marketing move on apple's part.

iJon

[zer0army]

Ok I have finally decided to register an account here after seeing these updates. I don't know if anyone mentioned this yet but I just though it was funny as hell.

On the main powermac page http://www.apple.com/powermac/ under the firewire800 paragraph at the end they decided to tack this little sentence on

" And it’s way faster than USB 2.0."

Then they just end it there not feeling the need to back it up or anything cause we all know it does truly kill usb2

I don't know I just found that really funny and though it was worthy of my first post.

:p :p :p

[MacFan25]

That seems like a great price for the low-end PowerMac.

[edenwaith]

Originally posted by iShater


So do you agree that the DP867 were a better choice than the 1Ghz low-end no?

Definitely. At first I thought the low end was a Dual 1 GHz machine, but when looking at it again, I was somewhat disappointed that it was only a single processor machine. Seems somewhat like a step backwards for Apple. Nice to see the lower price, but if I got a PowerMac now, I'd want at least a dual processor. It seems like if there is a new iMac with a 1GHz processor, Apple might be killing some of its lower end of the PowerMac line, depending on how desperate someone needs AirPort Extreme or the new Firewire 800.

Also, does anyone know if there are still the constraints on bus speed and the DDR RAM, or is the DDR RAM still somewhat limited on the bus speed of the machine?

[DeusOmnis]

This is all nice and dandy, but where is my iPod 2? BTW, i want it to be a PDA as well. So apple, since i know you read these forums, GET ON IT.

[sedarby]

Originally posted by trebblekicked
just a thought-
if the current motochip is stuck at 1ghz, and current 1.42 is a factory overclock job (which could lead to possible reliability issues), maybe the 7457 will make it's debut in XServe and iMac revisions, where ultra-reliability is more important...



Why would "ultra-reliability" not be important for PowerMacs?

[job]

Originally posted by DeusOmnis
This is all nice and dandy, but where is my iPod 2? BTW, i want it to be a PDA as well. So apple, since i know you read these forums, GET ON IT.

Arn't we a little self-centered today...

[job]

Originally posted by sedarby
Why would "ultra-reliability" not be important for PowerMacs?

The passage you quoted also leads to the question:

Why would Apple debut a new revision (read: faster/better) of the G4 in the iMac line so soon after speed-bumping the towers?

[Rocketman]

Originally posted by hitman
Just wanted to inject some humor in this thread:

My first brand new Mac was a 7200. It was a 601 running at 75Mhz.

I worked with that until I bought my second brand new Mac, a slot-loading iMac DV running at 400Mhz.

Keeping with the rate of Mhz increases when I buy a new Mac, I should purchase a new Mac when they hit 2133Mhz (see work below). :p :D

400/75 = 5.333
5.333*400=2133Mhz.

:)

Anyways, I don't think the 1Ghz is all that bad. Then again, coming from someone on a slot-loading iMac... :rolleyes:

2x1.25Ghz qualifies. Order today.

Rocketman

[Rocketman]

Originally posted by zer0army
Ok I have finally decided to register an account here after seeing these updates. I don't know if anyone mentioned this yet but I just though it was funny as hell.

On the main powermac page http://www.apple.com/powermac/ under the firewire800 paragraph at the end they decided to tack this little sentence on

" And it’s way faster than USB 2.0."

Then they just end it there not feeling the need to back it up or anything cause we all know it does truly kill usb2

I don't know I just found that really funny and though it was worthy of my first post.

:p :p :p

One thing it does NOT kill is the availability of USB2 devices that do NOT have firewire. Apple should support USB2 as a religious position. Now that they have FW800 they have nothing to fear, right? Hmmmm?

Rocketman

[MacKid]

Originally posted by eric_n_dfw

FALSE (at least partly)

Any multi-threaded application will automatically get a performance boost out of a dual G4 machine under OS X. OS 9 did not provide Symetrical Multi-Processing support so it was up to the application programmers to look for and utilize that 2nd processor. Now, if I start a 2nd thread in my code, OS X will run it on whichever processor it wants to. (That's a simplistic description, but basically accurate)

OS X can also automatically run two single threaded applications on separate processors thus giving better performance to both app's than if they were both running on one.

I think what he was saying is that although they might not get any slower, they won't get a speed boost with dual processors, simply because the threads weren't designed to be separated with parallel processing.

[MacKid]

Originally posted by smackdesign
So the powermacs are now at 1.4 ghzs. If apple updates the line every 6 months we should see them updated next in July. I hope they release the 970 at macworld NY and not another speed bumbed moto chip. I'm crossing my fingers.

You probably won't see it until early Fall.

[sedarby]

Originally posted by wumpus
The low-end represents a serious downgrade - the only reasons Apple did not retain th 867 at the lower price point is because they want to increase their margin on it while giving the illusion of better without cannibalising the iMac. The 1ghz SP is a price/performance DISGRACE. Nothing Apple can do about the FSB bottleneck, Motorola's poor fabbing or any of the G4s shortcomings (at least at this stage), so two processors, built-in Bluetooth, FW800, USB2, ATA/133 (if not Serial ATA) and 8x AGP are REQUIRED to give the ILLUSION at least of value.

When Ars Technica called the current 'Pro' lineup an 'overpriced embarrassment'
some months ago it was an accurate assessment. It is even more accurate now. The LCD and good looks of the iMac can compensate somewhat for PC-trailing performance. But for towers Apple needs to have dual CPUs across the entire line just to PRETEND to compete on performance - of course, at this stage informed pros who don't need a new machine NOW will wait until the 970 comes to close a bit of the gaping gap with WinTel. Maybe Apple knows this and is barely bothering? Why bother to add more than a speed bump when you know that this lead turkey architecture (the 'G4') is on its last legs?

Also, no-one seems to know yet if these are the 7457s or if there is full DDR support (somehow I doubt it). Are these the overclocked 7455s again? (before anyone gets upset about that word - Apple's chips are rated by Moto at that speed, just like I am sure there were 604e chips at the fab that could hit 400mhz with heavy chilling - but the 7455 DESIGN is NOT advertised by Moto as capable of more than 1ghz, thus Apple's are for all intents and purposes overclocked - thus the heavy heat sink and big noise).

And like others, I do get jealous of some of the advantages of commodity computing - just going onto Dabs.com every so often and get a CPU that will SMOKE any G4 for $70 is sickening sometimes?

..An Apple lover since the daze of the ][+?

A while back people where wishing for an entry level machine with slots. Well, the single processor 1 ghz machine is just such a machine. They upped the ante with a dual 1.42 ghz G4 and finally gave us Firewire 800. I'm very impressed with this offering.

[Macmaniac]

Man I wish some disposable income to get myself a new Powermac:) Maybe my school will solve that problem for me!

DROOL!!!

[Doctor Q]

Do we know how long it will take to get our hands on the new Power Macs and the 20" display? Are they available immediately or is there a scheduled shipment date?

[iShater]

Originally posted by Doctor Q
Do we know how long it will take to get our hands on the new Power Macs and the 20" display? Are they available immediately or is there a scheduled shipment date?

Hey Doc, do you update your signature everytime Apple updates the PM line? :D

[zer0army]

Originally posted by Rocketman


One thing it does NOT kill is the availability of USB2 devices that do NOT have firewire. Apple should support USB2 as a religious position. Now that they have FW800 they have nothing to fear, right? Hmmmm?

Rocketman

Wow buddy relax........:eek:

All I was saying was that I though it was funny the way they put it. I never said anything about availability of FW800 products I was just talking about its speed.

[dabirdwell]

Our 20" ships in 3 days.

Yup.

That's Right.

[dabirdwell]

But almost three weeks for my 12" powerbook. A week and a half for AE Basestations.

[janey]

i know i shoulda waited for the powerpc 970 but i couldn't wait...got a new 1.42ghz powermac from the apple store right now and the estimated arrival time is approximately a month ;)

[jrv3034]

Nice, Ubergeek!!!:D

The minute that thing arrives you MUST log on to MacRumors and tell us all just how droolworthy it is:cool:

We are all jealous;)

[janey]

dude it's übergeek, not Ubergeek.

Of course I will. But then I already know what it's like but sorry i can't tell you ;)

[Fender2112]

Originally posted by MorganX
Can someone please explain Apple's business model to me? As someone not interested in a PowerMac (I can trick out a Wintel box that blows it away for much less), why Apple feels they need to:

1) Lock their configurations
2) Keep performance speparation between the PowerMac Line and the iMac line?

Why doesn't Apple allow you to buy as much CPU as will fit in the iMac and all lines. Many home PCs are much more powerful that PC workstations. You order what you want. If I want a maxed out iMac, that doesn't mean if you cripple it, that I will then buy a PowerMac that I don't want.

I just don't get it.

Their product line up is actually quite simple:

......................Desk Top.................Lap Top
....Pro Line:...Power Mac.............Power Book
...Consumer:.....iMaci.......................Book

Each model usaully comes in three configurations that can usually be be tweaked to some extent.

I remember in the late 80's, early 90's. The product offerings were really confusing. You had models just for business, some for consumer, and others exclusively for education. And each came with it's own unique software bundle.

[Doctor Q]

Originally posted by iShater
Hey Doc, do you update your signature everytime Apple updates the PM line?Yes, although it doesn't have to be a Power Mac; any nifty new Apple toy will do. So far, you're the only one who noticed and mentioned it. I'm using small type so that nobody else will know the secret!

[raintalk]

Did they fix the fan noise? No sense making these bigger and faster if they start hovering around the room.

[sedarby]

Originally posted by hitman


The passage you quoted also leads to the question:

Why would Apple debut a new revision (read: faster/better) of the G4 in the iMac line so soon after speed-bumping the towers?

A better question is why would they put a better/faster processor in an iMac before the PowerMac? This goes along with the speculation I have read about Apple putting the IBM 970 in the iMac first. No way!

[DakotaGuy]

Originally posted by iShater


1) The iMac might not hit 1Ghz, they might go to 933 and keep the 800 as the low end, but I am not sure where that would leave the eMac.

2) They might, but they will sure do something to make it slower, especially now that the low-end PM has a single CPU.

3) I don't think it is worth the money. But that is my opinion, find a dual 867, it will probably be faster than the net iMac and you will get duals. This depends on how you plan to use the machine.

I still think a 1Ghz iMac/eMac is very possible. If anyone remembers back when the iMac was introduced the base PowerMac model was a single 800Mhz. The same processor speed as the iMac. Now I am not sure what will be done with the memory or bus speed, but I think Mhz will be matched.

I still predict that Apple will update both the iMac and eMac with either 867Mhz or 1Ghz processors. Right now they are focusing a lot of portable products on those two processors.

[TonicAngel]

They finally caught up with my PC in speed :) After I get my laptop, I'll go for a tower :D One more month until I can finally be a switcher :p

[Rajj]

I think Apple should have kept all the machines dual!!

You are getting raped of a semiconductor!!
:mad:

[Wry Cooter]

Originally posted by iAlan

But I am wondering, all those posters who are saying "YEAH!!", will they buy anything? It seems that half the posters here are always waiting for the next best thing and are all happy when it arrives, then a month later start complaining things aren't to their liking, and want yet another next best thing. Do these people have loads of cash to keep buying new stuff, and are these the same people who will wait for the 970, or do they not buy stuff at all but wish Apple had the best stuff on the planet anyway?

I will only say, this is what Apple should have done six months ago to meet reasonable expectations of MWNY announcements. The windtunnels fell short, they bombed. Does anyone have sales figures?

But now, is it six months too late? A lot of us have decided to wait for a 970, if that ever happens... We at least wanted the firewire 2 and Bluetooth on board, and a bus like the Xserve, which this release sort of at least has.

[Wry Cooter]

Originally posted by sedarby


A while back people where wishing for an entry level machine with slots. Well, the single processor 1 ghz machine is just such a machine. They upped the ante with a dual 1.42 ghz G4 and finally gave us Firewire 800. I'm very impressed with this offering.

Now if this offering had the processor ZIF upgradeable.....

[job]

Originally posted by Rocketman
2x1.25Ghz qualifies. Order today.

2133Mhz for one chip. :p :)

[stocke2]

does anyone think they will update the 15" powerbook soon?
I am planning on buying one(hopefully next week) and would hate to have them upgrade to 802.11g and a backlit keyboard and better ram a couple days after I order mine.

[bousozoku]

Originally posted by littlerich
I like to keep a track of the way that apple moves and I think its great that they do all this upgrading but when I get a new machine I might aswell get a pc. I have just bought a dual 1ghz and am pretty pissed off at the fact if I had waited until now then I could have got a dual 1.25 for the same price.. Or if I didn't get the dual 1.25 now just wait till August to get a dual 1.42 for the same price.. This never used to be like apple... If you buy an apple it usually keeps its value but the way thing are going it seems that apple may and possibly will lose alot of sales to pc's unless they make their prices as cheap as them and as cheap to upgrade. Anyone agree?

I'd like to agree but technology is moving too quickly and I'm just happy that Apple is moving along (with it would have been too strong.)

I have a dual 800 and, while I wish I could always have the fastest, I've resigned myself to the idea that it won't stay that way for long. My machine is now 3 upgrades behind the times. Oh well. It's still faster than my G3/400 and a good machine. Are the latest machines amazingly better? No.

[job]

Originally posted by stocke2
does anyone think they will update the 15" powerbook soon?
I am planning on buying one(hopefully next week) and would hate to have them upgrade to 802.11g and a backlit keyboard and better ram a couple days after I order mine.

I'd wait a little, depending on how soon you need it.

[stocke2]

Originally posted by hitman


I'd wait a little, depending on how soon you need it.


I am dying to move to a mac... I am really fed up with PC's, and as much as I like Linux for a server... OSX really is the premier UNIX for the desktop.
does anyone have an idea how long I should wait?

[Shrek]

Now that this has come out, I can really envision new iMacs running at 1.25GHz! Go Apple!!! :D Maybe now there might even be a 20" iMac if they use that new screen in it! Yay! :D :D :D

[ddtlm]

These replies go back a ways...

iShater:

So do you agree that the DP867 were a better choice than the 1Ghz low-end no?
Actually yes. :) I'm a SMP junkie. But I expect the single 1ghz to often be the faster machine.

iJon:

and the fx kills the 9700
Bzzzt! Well OK the FX can defeat the 9700, but at a higher cost in both money and noise.

edenwaith:

Also, does anyone know if there are still the constraints on bus speed and the DDR RAM, or is the DDR RAM still somewhat limited on the bus speed of the machine?
The FSB is still the bottleneck as far as the memory is concerned.

MacKid:

I think what he was saying is that although they might not get any slower, they won't get a speed boost with dual processors, simply because the threads weren't designed to be separated with parallel processing.
Threads are by nature separated for paralell processing. The thing is that most apps are not significantly multi-threaded, and by significantly I mean multithreaded in such a way that the single app can significantly benefit from having two processors.

raintalk:

Did they fix the fan noise? No sense making these bigger and faster if they start hovering around the room.
Yeah this is my question of the day too.

xrhajj:

I think Apple should have kept all the machines dual!!
Too many people probably got the dual-867 instead of the better models. Apple presumably wants to raise ASP's by moving some users up to the 2nd model, while gaining some more sales because of the lowered entry point. It may just do the trick for them.

stocke2:

I am really fed up with PC's, and as much as I like Linux for a server... OSX really is the premier UNIX for the desktop.
Damn straight! I wish Apple would match the command-line goodness of Linux, but I don't think that they are headed that way. Oh well...

[MacKid]

Originally posted by raintalk
Did they fix the fan noise? No sense making these bigger and faster if they start hovering around the room.

They claimed to have done it.

[Wry Cooter]

Originally posted by Le Big Mac


I'm missing this too. [why a single g4 entry model tower] Why not add bluetooth support and FW800 into the "old" Dual 867? Keep the price the same or drop it $100. For $200 less, one gets one less processor, and FW800. Not very impressive.:mad:

My guess it is due to educational customers, or others buying several macs at a time. They want access to the PCI slots if need be. They want an less expensive workstation for their cluster, where the heavy chewing is done at a dedicated tricked out workstation, and the rest are kept at 'good enough' level as far as their CPU goes, used for tasks that really don't need the horsepower (but are still machines allowed to talk to the other computers and drives as fast as Papa Bear Big Dog)

[LethalWolfe]

Originally posted by hobie


Oh boy, do you have any idea how the Dell system works? They have only 2 days of material in stock, demand their suppliers to deliver within 4 hours! And, believe it or not, they also have completely customisable laptops.
It is very possible that Apple has even more stock of raw material than Dell does.

Obviously not. :D

iJon Have you seen the Benches for the FX Ultra? Most gamers are calling it a flop because it barely beats out the 9700 (10% difference or less and it does lose to the 9700 in some benches). It's also big, hot, loud, power hungry, and expensive. The FX is slated to hit stores next month and ATi's next card is due out in March. And even if ATi's next card is only an minor improvemnt over the 9700 it will still kill an FX Ultra.


Lethal

[stocke2]

Originally posted by ddtlm

stocke2:


Damn straight! I wish Apple would match the command-line goodness of Linux, but I don't think that they are headed that way. Oh well... [/B]

the first thiing I will have to do is change to the BASH shell I am sure. However from what I have read about OS X and from playing with OS X on an ibook at the local circuit city(before they no onger sold macs, and before 10.2) it is one sweet system, but what would you expect...it is based off of NeXTStep

[wchamlet]

Whomever said that multi-threaded processors are faster than just one processor is wrong. At least not all the time. In Maya 4.5, I can render faster just using one processor than I can using two. Maya 4.5 is multi-threaded, and multi-processor enabled in rendering. However, using the test at www.highend3.com, I've found, along with other Maya users on the Mac, that dual processor Macs render faster just using one processor. I don't know why that is, but it's a fact none-the-less.

Actually, when I ran the test, it rendered just under four minutes using 2 processors. When I ran it again with one processor, it did it in less than 3. Which is kinda interestiing especially since it's a 1GHZ processor, when matched to AMD's or P4's, the G4 is actually faster at the same speed (mhz). Makes you think huh? Why can't they use Altivec and actually utilize the G4 for all it's worth?

[zer0army]

Has anyone done a good comparison of the new LCD’s against Formac’s ?

And what is the pixel response time on apple’s LCD’s ? they don’t seem to list it in the specs, they just say they have “Lightning-fast pixel response”.

What exactly is it though ?
:confused:

[Shrek]

One thing that I'm impressed with is that this time the high-end PowerMac is not $5,000. Way to go, Apple! :D

[MacKid]

Originally posted by Shrek
Now that this has come out, I can really envision new iMacs running at 1.25GHz! Go Apple!!! :D Maybe now there might even be a 20" iMac if they use that new screen in it! Yay! :D :D :D

Way too top heavy, buddy.:p ;) :D

[bentmywookie]

Ok, for those arguing about the single processor vs. dual processor issue, all I want to say is that an OS (at its lowest level) needs to do resource allocation/management. It's the interface between the hardware and the software.

So, to characterize the scenario, a piece of software comes along and tells the OS, "hey I need to cook this meal, here's the recipe, now off ya go!" And the processor looks and sees, "well I have two ovens, 3 bowls, 5 fridges, etc." and basically it uses whatever it has to get the job done. Ok, enough with this analogy (for now).

So if an OS like OS X is multiprocessor capable, it knows how to distribute its labor so that two processors can do it instead of one. Of course, there is some overhead to multiprocessing, and certain programs will be better suited to multiprocessing than others. But all of this should be invisible to the software.

I'm not familiar with Mac programming though so I'm not sure if there are ways to code "hints" to the OS to indicate that these areas are optimized for MP or not.

But basially, that's how it works. So intuitively, you would think MP is always better, but its not necessarily so due to overhead and the nature of the task.

Also - to the person who said most apps are single threaded - what is your justification for that?

[ktlx]

Originally posted by Dont Hurt Me
Then get a dualie, but as far as mac gaming goes most ports( read ports) are usually not using all this cool mac stuff nor the 2nd processor but some of the better companies try to make use of it such as id, But when i see a great new title like Unreal tournament 2003 will not use SMP then of coarse i want clock cycles. Remember Apple did all that dualies stuff to try to make up for where motorola was not taking them! And to a large extent they have succeded with super cool OSX. Even now though it looks like it is still up to them since no g5 or 7457 from motorola! MOTOROLA sure is quite about their CPU's

I am sorry, but if your total focus on Macs is games, you are a doofus. You are paying way too much money for way too little. If you cannot stand something by Microsoft then buy a PS.

Apple should be creating PowerMacs that address its market strengths. Those are in content and media creation. Those areas benefit the most from dual processors because the applications are multithreaded.

[ktlx]

Originally posted by iJon

and the fx kills the 9700. and so on and so on. both companies will always be a step behind each other, it just depends if your talking to a ati or nvida person. on my mac i could care less who makes my video card but on my pc its nvidia all the way, ati has poor driver implentation on the pc said and the ti's and fx's beat em, wouldnt have anything else. now lets get those audigy drivers out for mac and our machines will kind of be gaming machines.

Read the benchmarks. The ATI 9700 smokes everything from nVidia but the 5800 Ultra and even those results are mixed. Some places find the 9700 slightly faster some places find the 5800 Ultra slightly faster. My guess though is that once nVidia has a chance to tune their drives, they will beat the 9700 in all of the benchmarks.

If you want to read more about the new nVidias versus the 9700, go to ArsTechnica and follow the links they assembled. There are about five or six comparison reviews.

[ktlx]

Originally posted by bentmywookie
Ok, for those arguing about the single processor vs. dual processor issue, all I want to say is that an OS (at its lowest level) needs to do resource allocation/management. It's the interface between the hardware and the software.

So, to characterize the scenario, a piece of software comes along and tells the OS, "hey I need to cook this meal, here's the recipe, now off ya go!" And the processor looks and sees, "well I have two ovens, 3 bowls, 5 fridges, etc." and basically it uses whatever it has to get the job done. Ok, enough with this analogy (for now).

So if an OS like OS X is multiprocessor capable, it knows how to distribute its labor so that two processors can do it instead of one. Of course, there is some overhead to multiprocessing, and certain programs will be better suited to multiprocessing than others. But all of this should be invisible to the software.

I'm not familiar with Mac programming though so I'm not sure if there are ways to code "hints" to the OS to indicate that these areas are optimized for MP or not.

But basially, that's how it works. So intuitively, you would think MP is always better, but its not necessarily so due to overhead and the nature of the task.

Also - to the person who said most apps are single threaded - what is your justification for that?

In a word, not even close.

Writting an application to take advantage of more than one processor is not invisible to the software. The software has to be specially written to break up the work in a way appropriate to the application. If you want to optimize for a dual processor system (and there is no reason not to since there are no quad processors Macs), you create two worker threads and try to split the work among the two threads as best as you can.

[iJon]

Originally posted by LethalWolfe


Obviously not. :D

iJon Have you seen the Benches for the FX Ultra? Most gamers are calling it a flop because it barely beats out the 9700 (10% difference or less and it does lose to the 9700 in some benches). It's also big, hot, loud, power hungry, and expensive. The FX is slated to hit stores next month and ATi's next card is due out in March. And even if ATi's next card is only an minor improvemnt over the 9700 it will still kill an FX Ultra.


Lethal
well from the benches i have seen the 9700 barely beats out the ti4600, so it depends how you look at it. plus many more games these days are being coded to perform better on the nvidia side. also, "hot,power hungry, and expensive" all i can say is that we have powermacs dont we. the fx has got ddr2 and a faster bus than the 9700, .13 micron and in all games i have seen the fx beating the 9700, but the higher rez's you go the less it shines.. I am no video card expert, thats why i dont like to get in arguements about it because i probably dont know what im talking about (i like mac arguements because i am a mac expert. :) ) but just from what ive read the fx looks pretty promising. they are both very wonderful cards but i choose to use nvidia because ati has poor drivers for the windows side and more and more programs are being written to take advantage of nvidias techs.

iJon

[superfula]

Originally posted by iJon

well from the benches i have seen the 9700 barely beats out the ti4600, so it depends how you look at it. plus many more games these days are being coded to perform better on the nvidia side. also, "hot,power hungry, and expensive" all i can say is that we have powermacs dont we. the fx has got ddr2 and a faster bus than the 9700, .13 micron and in all games i have seen the fx beating the 9700, but the higher rez's you go the less it shines.. I am no video card expert, thats why i dont like to get in arguements about it because i probably dont know what im talking about (i like mac arguements because i am a mac expert. :) ) but just from what ive read the fx looks pretty promising. they are both very wonderful cards but i choose to use nvidia because ati has poor drivers for the windows side and more and more programs are being written to take advantage of nvidias techs.

iJon

The thing you are missing is that the FX "LOOKS" promising. But it's a flop. Much more money for something that doesn't provide that much performance increase. Like was said earlier, all the current accurate benchmarks has the FX barely beating the 9700, and in some tests, the 9700 wins.

BTW, ATI drivers aren't crap anymore. I'd take them over Nvidia anyday. Nvidia may roll out new drivers, but it won't increase the performance enough to warrant the huge price difference.

The FX is dead even before it's released.

[SoonToGetAMac]

Did anyone else notice the irony that on the Apple.com main page, the LCD is shown with a Space Shuttle taking off, and that today is the 17th anniversary of the explosion of a Space Shuttle as it was taking off.

[bentmywookie]

Originally posted by ktlx


In a word, not even close.

Writting an application to take advantage of more than one processor is not invisible to the software. The software has to be specially written to break up the work in a way appropriate to the application. If you want to optimize for a dual processor system (and there is no reason not to since there are no quad processors Macs), you create two worker threads and try to split the work among the two threads as best as you can.


Nor did I say this - it IS invisible to the software though whether or not the OS uses one or two processors. That's why Photoshop (which people say is MP-enhanced) runs on DP Powermacs and iMacs.

In the time it took you to write this cheeky response, you could have reread my post and seen that you would not refute anything I said.

[kansaigaijin]

correct me if I am wrong, but are these not Xserve boards (technology) in Powermac cases?
they have the new I/O bus etc from the xserve.

nobody read the tech spec pages? before spouting off? That would explain why they do not boot OS9 and why the low end model is still an improvement with only a single processor.

It also explains why there is a link to OS9 bootable machines using the old technology.

and don't bother feeding the trolls, they will never bother to go read the tech sheets, and wouldn't understand them if they did.

[pyrotoaster]

Originally posted by stocke2
I am dying to move to a mac... I am really fed up with PC's, and as much as I like Linux for a server... OSX really is the premier UNIX for the desktop.
does anyone have an idea how long I should wait?
Wait as long as you physically can.
If a SuperDrive 12" AlBook doesn't suit your tastes, and you can't afford the 17" (you aren't alone if you can't!), just wait.
15" Powerbook upgrades are in the near future. The advantages of a new Powerbook, like Firewire 800 (even if it isn't on the 12"), Airport Extreme, an Aluminum case, a possible backlit keyboard, a faster processor and SuperDrive, and DDR RAM, are worth the wait.
Keep an eye on those 15.4" Powerbook rumors (the current ones are 15.2", suggesting a minor case redesign, possibly just a side-effect of a stronger Aluminum case), as they should point to a release time.
I don't think new 15" PBs will be announced until the new ones ship (at least the 12" model), so a mid to late February/March time frame isn't unrealistic. I hope you can hold out that long.

BTW, I hope nobody bought a 23" Display yesterday! The new LCD prices are great! :)

[iAndy]

Before I specifically respond to some comments, I just want to say that:
i) I own a number of PCs (I know - boo hiss !)
ii) I now own two Macs (soon to be 3), and have bought more for work
iii) This thread is neither intended to Apple bash or Wintel defend, just try and put some things in perspective from someone who uses both
iv) I still hold shares in Apple and Intel ;)

jettredmont
Okay, hold on a minute. Are you talking about upgrading your CPU? I can tell you that in 12 years of owning PCs I've only attempted to upgrade a CPU once (which was a failure), and have never otherwise seen any benefit in doing so. Sorry this doesn't prove that PC are difficult to upgrade, only that you might not be too hot recognizing one end of a screwdriver from another ;)


jettredmont
I've always found that when I start wanting to upgrade my CPU it is cheaper and easier to just replace the whole box and relegate the old box/CPU to the basement server pasture. That is not necessarily the case. There are a number of companies out there making a very healthy living from cpu upgrade cards. My 14 year old nephew recently upgraded my 3 year old Dell XPS-T500 (still going strong) from 500MHz to 1.3GHz for ~$120. It arrived, he took out old processor, slotted in upgrade processor card, ran suppled driver disk - the whole process took him less than half an hour, and couldn't have been much easier.


jettredmont
Back to upgrading a CPU. First, "$200" is not the price of your average CPU upgrade. Unless you bought a behind-the-curve CPU to start with, the motherboard you currently own will generally not accept the latest CPU.
WRONG - sorry your PC knowledge appears to be way out of date (read below)


jettredmont
You have to also replace the motherboard.
NOPE


jettredmont
Which often (in my experience, at least, from DIMMs to FPO to EDO to SDRAM to DDR to RAMBUS to DDR 2 ...) means replacing memory, which of course isn't as expensive as it used to be but it still doesn't come free. Nowadays you'll also likely have to add in a new fan/cooling system.
WRONG AGAIN - most upgrade processors come with additional fans/heat sinks built in


jettredmont
Take a new CPU at $200 and a new motherboard at $150 and a new gaggle of memory at $200 and you're quite a bit beyond what Mac users pay when they want to upgrade their CPU (with a daughtercard CPU that works on their existing motherboard). Granted, your $550 gave you the latest motherboard features as well, which is a nice side-effect, but in my experience working with PC upgrades, the $200 CPU upgrade is pure, unadulturated myth.
Boy are you digging a hole for yourself here ! ;) Check out www.Powerleap.com for info on their latest 1.4GHz upgrade offering @ $160 - enough said !


jettredmont
And don't just take my word for it. Look into it at Tom's Hardware and cNet/ExtremeTech and Ars Technica. It's rare to see someone arguing that a CPU upgrade is financially sound, and then it is with caveats such as "your current MB will accept the latest/greatest" and "you are adding 25%+ of frequency to your CPU".
No I won't take your word for it, because there are numerous articles giving detailed info on PC systems which are worth upgrading. Furthermore Powerleap (an unheard of Scandinavian outfit just a couple of few years ago) has already sold over 500,000 upgrade CPUs on its own, and in their own words have barely scratched the surface !


jettredmont
Hmmm. So I guess Dell is doing something horrendously wrong with its 2% of the PC market?
WRONG - I think someone has already corrected you on Dells current market share


jettredmont
Apple does not need to dominate the PC industry in order to succeed. It is doing quite well currently, reporting ongoing profits quarter after quarter.
Actually, as an Apple shareholder, I was sorry to see that Apple actually made a loss last quarter. So this is certainly no time for Apple (or its users) to become complacent ! Personally I am very excited about the direction that SJ is currently taking the company, so much so that I have increased my AAPL shareholding since MWSF.


Freg3000
I am going to make this very very simple. I don't care how cheap or expensive you can buy a PC for. The fact is, PC's cannot run iTunes, iMovie, iPhoto, and iDVD.
Period.
So that's a no to buying PCs then ;)

Actually you help make a very good point Freg3000, and surely that is that once we get through all the marketing hype and froth computers are just tools - and NOT religious icons ! Although that distinction isn't always easy to see from some comments made by the more extreme pro-Mac and pro-Wintel camps. But at the end of the day it comes down to who provides the best tools for each end users job. And IMHO the answer to this question today is not simply black (Wintel) or white (Apple) !

Currently I believe that Apple provides the better multimedia solutions (even with lower CPU speeds), especially with movie editing and DV burning. That is why I have bought a superdrive iMac for home movies and ordererd eMacs for our hospital.
The recently introduced iLife upgrade looks like it will keep Apple ahead of the pack in this area.

But on the otherhand, for simple word processing, spreadsheets, database development, web browsing etc. I believe it is much harder for Apple to currently compete head to head on price performance with quality 3GHz+ Intel boxes from companies such as Dell.

However Apples's move into the DigitalHub looks very attractive, although there appears to be a lot of work still to be done until the iSync/iCal apps are as polished as iTunes. But I believe that Apple's great track record of innovation (vs Microsoft) could open up a huge market if they can execute fast enough in this area. Unfortunately articles such as PCMagazine recent "First Look: A Mixed Bag of Apples"
http://www.pcmag.com/article2/0,4149,847259,00.asp are not going to help Apple win over possible switchers. Actually I was very p#ssed off to read this article, as it seemed as biased against Apple as a few folk here are against W#ndows (but who haven't tried using a quality brand PC with the latest version of XP Pro preinstalled).

With the introduction of OS-X's improved connectivity with Jaguar, I believe (and hope) that more and more PC users are going to start buying Macs as PART of their home or office network solution - but will probably never completely "Switch" over to Macs until price/performance reaches near parity with Wintel equivalent. There again if Apple can get their G5 rolled out asap, they may still win the day ;)

Sorry for the long post folks, but it is really a cumulative response to numerous myopic Apple and Wintel bashing threads.

Anyway back to the thread title - GREAT NEWS ON THE DISPLAY PRICES APPLE - now that's how you win over more switchers !!!

P.S. and where is my PB 12" (SD) ;)

Andy

[yosoyjay]

Originally posted by Nebrie
If you think Apple can serve the pc customization market that overclocks and watercools AND maintain quality, you must really be high. That is the dumbest thing I have ever heard.

If you are referring to my quote where I asked if Apple was water cooling their chips it was a joke. I said it in the context that the 1.42GHz must be an overclocked 1.0GHz which would most likely need some serious cooling. I will refrain from futher comment.

[superfula]

Originally posted by kansaigaijin
correct me if I am wrong, but are these not Xserve boards (technology) in Powermac cases?
they have the new I/O bus etc from the xserve.

According to Apple, you are correct
http://www.apple.com/powermac/

[ddtlm]

wchamlet:

At least not all the time. In Maya 4.5, I can render faster just using one processor than I can using two.
I'd have to assume that main memory contention is the problem.

Which is kinda interestiing especially since it's a 1GHZ processor, when matched to AMD's or P4's, the G4 is actually faster at the same speed (mhz).
Actually, without AltiVec a G4 is probably slower per clock than an Athlon. The Athlon core is much more advanced.

Why can't they use Altivec and actually utilize the G4 for all it's worth?
Well from experience I have to say that vector instructions take extra planning and extra programming, and even if someone did use AltiVec really well, they might just find themselves bottlenecked on memory accessed. Varies a lot depending on the task at hand of course.

Shrek:

One thing that I'm impressed with is that this time the high-end PowerMac is not $5,000. Way to go, Apple!
$5000 gets you one spiffy Mac with 23" LCD. Very nice.

bentmywookie:

I'm not familiar with Mac programming though so I'm not sure if there are ways to code "hints" to the OS to indicate that these areas are optimized for MP or not.
Just like (most) AltiVec optimization, taking advantage of multiple CPUs (threading or forking) takes extra planning and extra programming. From experience, it is very possible to try to thread something and end up slower than you started. :) Threading is just like AltiVec in that it has some places it can be used, some places it can't, varying milage everywhere, and is generally a lot of extra work. Sometimes it's worth it, sometimes it's not.

Also - to the person who said most apps are single threaded - what is your justification for that?
Well other than the fact that programmers won't invest that effort unless they have to, I think that it is pretty easy on a dual CPU machine to see what apps are and what apps are not threaded. Here's how it is:

For the sake of completeness, I'll say that threading has two uses that I see, one is to increase application responsiveness by allowing input to continue while the app works, and the other is to speed up computation. The responsiveness one is not very interesting because its pretty easy to do with our without threads, and since only one thead is doing work anyway, it works fine on a single CPU machine. No benefit to a dual CPU machine here, unless taking mouse presses somehow is bogging down your CPU. :)

Anyway, the other use of threading is to get more work done by sending work to two processors. This should be easy to spot. First of all, there needs to be a fair amount of work to do, so forget an email program or ICQ or whatever, not meaningfully threaded for speed. If there is a lot of work to do, I ought to be able to fire up 'top' or some CPU monitor and see the app use both processors, even if for just a little while (else there's no real extra work being done so who cares about threading). So I do this, and I see that Mozilla doesn't use threads for speed when rendering pages. Other apps like games are generally known to be one way or the other, Quake3 can use two chips, WC3 can not. I can see this on 'top' or whatever tool. And so on... you can do this for every app you have and you'll find that most all of them are not using more than a single CPU. They may seem to be using 50% of both CPU's, but thats just the OS moving it around really fast. Look for an app using well over 50% of both for extended periods of time. Anyway, I've never spotted any common app doing that, not in OSX, not in Linux, not in Solaris or Irix... nowhere.

About the only hole in my argument is the possiblity of threading on bits of work so small that they flash by before the CPU monitor or anything else mentions them to me, and so infrequent that my odds of ever catching one in the act are small. I guess you can choose to believe this if you want, but I'd have a hard time imagineing anyone taking the time to thread anything like that, and in any case, I can't imagine that such a thing would benefit much from a dual-867 over a single 1000. Threads do have overhead, you know, and such a thing would actually be slower in any situation where only one CPU was available (two threads on one CPU == silly), even if the system was a dual, but was perhaps running something else on the other chip.

Anyway that probably ended up overly long and complex. :(

ktlx:

In a word, not even close. ...
Good response, says I.

bentmywookie:

In the time it took you to write this cheeky response, you could have reread my post and seen that you would not refute anything I said.
Ummm... I think he did address your following comment:

I'm not familiar with Mac programming though so I'm not sure if there are ways to code "hints" to the OS to indicate that these areas are optimized for MP or not.

kansaigaijin:

and don't bother feeding the trolls, they will never bother to go read the tech sheets, and wouldn't understand them if they did.
Eh? Perhaps if you had been tuned in earlier you would have seen Apple talking about using Xserve tech on the MDD's this past fall, which ran OS9 fine.

[iShater]

Originally posted by stocke2



I am dying to move to a mac... I am really fed up with PC's, and as much as I like Linux for a server... OSX really is the premier UNIX for the desktop.
does anyone have an idea how long I should wait?


Go to an Apple Store and play with teh 12" PowerBook, if I had the money and Apple still had the 6 month 0% financing, I would have walked out with one. Those baby are sweeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeet! :D :D :D

I mean it! The pics of the machine do NOT do it justice, you have to see it and touch it to really appreciate it. Now, off to an Apple Store you go! go go go! :p :D

[stocke2]

Originally posted by iShater



Go to an Apple Store and play with teh 12" PowerBook, if I had the money and Apple still had the 6 month 0% financing, I would have walked out with one. Those baby are sweeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeet! :D :D :D

I mean it! The pics of the machine do NOT do it justice, you have to see it and touch it to really appreciate it. Now, off to an Apple Store you go! go go go! :p :D

I am sure it is nice...but is just not enough for me. the 17" is really nice, but shipping time too long, and is a little larger than I want... 12" too small 17" too big...15" really nice. ;)

[iShater]

Originally posted by stocke2


I am sure it is nice...but is just not enough for me. the 17" is really nice, but shipping time too long, and is a little larger than I want... 12" too small 17" too big...15" really nice. ;)

then get the 12" PB, get a bookendz dock, a nice 17" Apple display, keyboard, mouse, a viola! ;)

The 15" is guaranteed to be upgraded to match the other PBs, it is probably only a matter of time (when the systems in the distirbution channel are out). They are also probably doing the same thing they did for PMs, wait as long as they can to get as many sold as possible. So if you go for the current 15", you will be back here in a few months saying "man, I just got this!".

You can always resell the 12" if it is not good enough! :D

[zimv20]

Originally posted by Inhale420
'powermac' and 'value' still don't belong in the same sentence.

yet there they are

[eric_n_dfw]

Small threads are used all over the place. You seem to think that coding them is in some way difficult. It is not. You code your objects to be thread safe (fairly trivial if you know what you are doing), you instantiate one and tell it to run. Sometimes you create a bunch and (in Java terminology) put them in a thread pool, ordering all of them to fire off at the same time.

I've never looked at Mozilla's code but if you think about it, threads would be a obvious way to pull in images for an html page. You fire off 1 for each image and the independantly load the graphics without the main thread worrying about them.

I'd presume that Final Cut Pro does similar type things - firing off a pair of threads to render one frame each. Now - special code may exist to tell Final Cut Pro how many threads to issue at a time, there is a startup script that asks the OS what speed and number of processors exist. It uses that data to estimate how much real-time rendering power it should present to the user. It then can just fire off those threads at will, OS X puts them on the next available G4 time slice. This also allows a yet-to-be invented 4,6,8,etc processor machine to also be utilized to it's full extent.

Don't believe me about the browser point? Try OmniWeb and open the inspector that shows the page rendering status (I can't remeber it's actual window name) - each element of the page shows up and reports it progress - you can kill any one element. This is multithreading.

Now, I'm no game developer, and know nearly nothing about the craft, but I'd assume that the reason UT on x86 is so much faster than on a dual G4 is that the programmers have tuned it to that hardware much more - puting a lot more attention to any bottlenecks. I'd be willing to bet that hand tuned, assembly language is written in those 3d engines. Porting that to PPC and getting it to run as fast would be a major undertaking. They probably just take the c or c++ code and recomplie in with little or no asembly tuning. (Like I said, this is wild-arsed guessing ;) )

[bentmywookie]

Originally posted by eric_n_dfw
Small threads are used all over the place. You seem to think that coding them is in some way difficult. It is not. You code your objects to be thread safe (fairly trivial if you know what you are doing), you instantiate one and tell it to run. Sometimes you create a bunch and (in Java terminology) put them in a thread pool, ordering all of them to fire off at the same time.

I've never looked at Mozilla's code but if you think about it, threads would be a obvious way to pull in images for an html page. You fire off 1 for each image and the independantly load the graphics without the main thread worrying about them.

I'd presume that Final Cut Pro does similar type things - firing off a pair of threads to render one frame each. Now - special code may exist to tell Final Cut Pro how many threads to issue at a time, there is a startup script that asks the OS what speed and number of processors exist. It uses that data to estimate how much real-time rendering power it should present to the user. It then can just fire off those threads at will, OS X puts them on the next available G4 time slice. This also allows a yet-to-be invented 4,6,8,etc processor machine to also be utilized to it's full extent.

Don't believe me about the browser point? Try OmniWeb and open the inspector that shows the page rendering status (I can't remeber it's actual window name) - each element of the page shows up and reports it progress - you can kill any one element. This is multithreading.



Thank you - couldn't have said it better.

ddtlm - your only possible hole about having threads so small that they fly by you on the CPU monitor is a pretty big hole as per the reasons above.

[eric_n_dfw]

Originally posted by bentmywookie


Thank you - couldn't have said it better.

ddtlm - your only possible hole about having threads so small that they fly by you on the CPU monitor is a pretty big hole as per the reasons above.
Years ago, when I was a pretty green programmer, I thought threading sounded complicated. Once you've done it, it can solve a lot of performace (or percieved :) ) performace problems.

Here's a real-world example. I wrote a simple JSP based app that would pull info off of Yahoo's NFL pick'em pages and compile all of my friends football picks into a nice table. Originally, it would loop through all 30 players, hitting the Yahoo! URL for each player and waiting for the response. To speed it up, I wrote a single Java class that did this for a single player. Then, in my main program I created a ThreadPool, created 30 instances of that new class, telling each one to "run" as I created it and added each one the the pool.

Then I waited for the pool to report that all of it's threads had died (or timed out). Then I could ask each object instance for the data it had stored.

Threads are great for IO bound things like that - fire and forget if you will.

[eric_n_dfw]

Here's a link to the window in OmniWeb (http://www.omnigroup.com/images/applications/omniweb/DocumentInfo.jpg) that I was talking about.

[ddtlm]

eric_n_dfw:

You seem to think that coding them is in some way difficult. It is not.
I have coded things with threads on several occasions, and I've also done things with forks. Threads can be very difficult to apply.

I've never looked at Mozilla's code but if you think about it, threads would be a obvious way to pull in images for an html page. You fire off 1 for each image and the independantly load the graphics without the main thread worrying about them.
You're trying to tell me how things are, yet clearly did not underand my post: these (hypothetical) threads are not CPU-intensive. They spend 99% of their time waiting. Category #1: works just as well on a single-CPU machine.

In any case, this setup is probably not realistic because of the way networking works. Now, I'm not real up to date on the HTTP protocol, but it rides on top of TCP/IP and TCP/IP allows the browser to send off a bunch of packets (a HTTP request in this case) without needing to wait for return packets (HTTP data in this case). The returned data (web page, image or whatever) will pile up on the computer's buffers, and when the app goes to check on it the data is presented to it. This means that threads are not needed to endlessly poll for return data; a single thread can check at it's convienience. Mozilla's default mode actually is serial, one request after another and it does nothing till it has the previous bit of data. There is some other mode that they might call request pipelining that sends out lots of requests at once and theoretically boosts speed on fast networks. That fast mode can also be done without threads, because (if multiple sockets are needed) you could simply have, ta-da, multiple sockets that the browser writes requests to and checks for input from (when convienient). There is no reason that I can think of right now that this would benefit from threading, since I bet only one socket can actually be written to or read from at once anyway. Additional threads would spend most of their time sitting around doing nothing, so why not just do all in one thread?

As I've gotten more threading/networking experience, I actually have gotten less inclined to use threads.

I'd presume that Final Cut Pro does similar type things - firing off a pair of threads to render one frame each.
This falls into category #2, where substantial work is being done and I can probably see it happen in 'top' or something. I never claimed that multithreaded apps do not exist, I've simply claimed that they are not the common apps.

Don't believe me about the browser point? Try OmniWeb and open the inspector that shows the page rendering status (I can't remeber it's actual window name) - each element of the page shows up and reports it progress - you can kill any one element. This is multithreading.
I've never used that app and really can't comment either way. You could be right about it, or not.

bentmywookie:

ddtlm - your only possible hole about having threads so small that they fly by you on the CPU monitor is a pretty big hole as per the reasons above.
Sheesh, does noone understand what I wrote?

eric_n_dfw:

Years ago, when I was a pretty green programmer, I thought threading sounded complicated. Once you've done it, it can solve a lot of performace (or percieved) performace problems.
Tell me when you realize that you don't need threading for most of it. :)

[eric_n_dfw]

Just in case this horse is not dead yet -- in the Terminal, run "top"

The "#TH" column shows how may threads the process is using.

ps -M will show threads and the stats for each process as will

[ddtlm]

eric_n_dfw:

Here's a link to the window in OmniWeb that I was talking about.
Yes but why do you claim threads are involved?

[ddtlm]

eric_n_dfw:

Just in case this horse is not dead yet
No, not dead.

The "#TH" column shows how may threads the process is using.

ps -M will show threads and the stats for each process as will
Yes, but which are category #1, and which are #2? The existence of threads means very little, you still have to prove that there is significant work being done in paralell.

[pgwalsh]

Originally posted by SoonToGetAMac
Did anyone else notice the irony that on the Apple.com main page, the LCD is shown with a Space Shuttle taking off, and that today is the 17th anniversary of the explosion of a Space Shuttle as it was taking off. Did you remember that off the top of your head? If you did, then you have one hell of a memory.

[eric_n_dfw]

Threads can be difficult - usually, when they are, a better solution and/or design is called for anyway.

You are correct about the fact that IO processing is outside the scope of your original arguement - sorry, I get side tracked easily ;)

Just curious - what language and/or platform do you have your experience in? My experience (with threading at least) is from C++ on Solaris (with RogueWave tools) and Java on Solaris, OS X and Linux.

Admitedly, I've never had the "pleasure" to try threading in C.

I particularly enjoy Java's threading support (and am looking forward to getting time to learn Objective-C/Cocoa's implementations) as it does all the hard work for you.

[ddtlm]

eric_n_dfw:

My experience (with threading at least) is from C++ on Solaris (with RogueWave tools) and Java on Solaris, OS X and Linux.
Excellent choices. :)

I've got a few years of C/C++ with threads (posix pthreads) and networking via the BSD sockets. I've played with them in their SDL (http://www.libsdl.org) form a little in order to examine Windows compatible threading and networking. Most of my forking and creative IPC was done in Perl, I've done some Java and a fair amount of PHP, some Tcl, some x86 assembly, and who knows what else. The heavy stuff has been C/C++ and Perl though.

I've done most of my programming in Linux, but I use OSX more these days for the SDL/C stuff because I like project builder a lot (which is one of those apps that does use 2 CPUs for performance). I program in Windows enough to make sure my stuff compiles there. ;) Edit: I've also done some programming on Solaris and Irix, but not a whole lot. Seemed a lot like Linux except with more stupid problems.

More to the original thread, I've been really tempted to spend $5k on a new dual 1.42, combo drive, 512mb, 120gig, 23" LCD, rad-9700 system. Drool. But what would I do with my current system? :(

[eric_n_dfw]

Originally posted by ddtlm
eric_n_dfw:


Excellent choices. :)

I've got a few years of C/C++ with threads (posix pthreads) and networking via the BSD sockets. I've played with them in their SDL (http://www.libsdl.org) form a little in order to examine Windows compatible threading and networking. Most of my forking and creative IPC was done in Perl, I've done some Java and a fair amount of PHP, some Tcl, some x86 assembly, and who knows what else. The heavy stuff has been C/C++ and Perl though.

I've done most of my programming in Linux, but I use OSX more these days for the SDL/C stuff because I like project builder a lot (which is one of those apps that does use 2 CPUs for performance). I program in Windows enough to make sure my stuff compiles there. ;) Edit: I've also done some programming on Solaris and Irix, but not a whole lot. Seemed a lot like Linux except with more stupid problems.

More to the original thread, I've been really tempted to spend $5k on a new dual 1.42, combo drive, 512mb, 120gig, 23" LCD, rad-9700 system. Drool. But what would I do with my current system? :(
Do you do much OO? The reason I ask is that it seems that threaded designs and good OO design kind-of goes hand in hand. While procedural programming would require you to do a lot of "roll your own" coding to support threading. (I could be wrong)

I'll tell you one application that needs a lot of work - whether it be multi-threading or just plain re-writting, is Quicken. That thing is such a slug for simple things that it's embarasing!

As far as the new machines - I'll tell ya', the Student Developer Discount price on the top-o-the line with 512MB and the 9700 ATI is not much more than I paid for my G3/400 B&W back in '99!! :D (no discount for me back then - my company is putting back through college now at 31 :D) so I am mighty tempted!!!

[barkmonster]

All this talk about threads reminded me of something.

Does anyone remember all those excited reviews when OS 8 came out and supported multithreading ?

People were speculating about loads of uses for theads, the most useful one for me was the idea that photoshop could be coded to use threads so you could fire off multiple filters on multiple documents at once while scanning in a new documents and printing another. It never happened and photoshop still works like some kind of monotasking graphics OS.

The one area where threads could have been used under OS 8 - 9.x would have been when printing, even with background printing enabled I still stare at a model dialogue box for a number of minutes before documents are even added to the printer que.

There's numerous areas where threads could have been using for years on end but no one used them.

Anyway back to the new 20" displays...

They're OS X only!!!

not fair!

boo hoo!

[Sol]

Your idea of a multi-threaded Photoshop sounds similar to batch-actions (you know, converting a folder of images to another format or something like that).

OS 8 and later 9 did benefit from this multi-threading. The control strip for one did a great job of running alongside any hardware-intensive application. Audio CDs played non-stop even when rendering movies with QuickTime Pro. I am sure you can come up with your own examples.

For its day, and for any single CPU G3 Mac, OS 8 and especially OS 9 were very good OSs. The move to a UNIX-based OS was not what I expected it to be but let me just say that for a multi-processor system this is the only OS I would want to use.

[eric_n_dfw]

The problem with printing in Mac OS Classic has little to do with threading and much to do with the lack of premptive multitasking.

If photoshop took advantage of multi proc's in OS 8/9, then it was multi-threaded or at least spawned processes.

[MacQuest]

Originally posted by raintalk
Did they fix the fan noise? No sense making these bigger and faster if they start hovering around the room.

OMG!!

You made my breakfast OJ burst out of my nose imagining that!!

LOL!!:D

[Sol]

If you are reading this and eating breakfast at the same time then I can imagine what you are using as a cup holder for your coffee.;) Good thing not all Macs are slot-loading yet.

[Dont Hurt Me]

Originally posted by ktlx


I am sorry, but if your total focus on Macs is games, you are a doofus. You are paying way too much money for way too little. If you cannot stand something by Microsoft then buy a PS.

Apple should be creating PowerMacs that address its market strengths. Those are in content and media creation. Those areas benefit the most from dual processors because the applications are multithreaded. Another boo--- hooer---you use your mac for gaming! you cant do that,you shouldnt do that and yada yada yada!1 more time i use my mac for everything and even gaming believe it or not. thank goodness someone at apple has realized this have you seen there gaming page. oops i forgot your not supposed to do that and get back to that gauzeeeeeen bluuuurrrrrrrrr!The mac can should and will do all period!After all it is the digital hub. and with that i cant wait untill UT 2003 and DOOM3 is released so take that all you non gamers who are missing so much fun on the MAC. Gaming is a great way to compare machines, video cards, etc. If this was not so mike would not be using those figures all the time included in his tests at that great site Accelerate your Mac! Iam sure there are a lot of people who would love to see a 1.25 or 1.42 in a new imac but we wont probably get that because again it would eat into the pro line and all those non gamers might buy 1 of those machines instead of a powermac!The new displays are cool! the powermacs are moving forward!I just wish they would let the imac soar thats all!

[Trekkie]

Originally posted by appeLappe
I?ve noticed that you can choose to update the GPU to either the Geoforce4 Ti or the ATI 9700 Pro for the same price.
[/B]

That for me is a no-brainer. The 9700 Stomps the GeFroce4 Ti in all ways right now. For the same price, the 9700.

[barkmonster]

Your idea of a multi-threaded Photoshop sounds similar to batch-actions (you know, converting a folder of images to another format or something like that).

An example of a batch action :

Image1.tiff -> play action -> save file -> Image2.tiff -> play action -> save file etc...

This is linear, not parallel, you can't do anything at all with photoshop apart from leave it in the background and use another application while it's batch processing. Every stage of a batch process comes after the previous one, you can't run 10 seperate batch processes at once for instance. Whether it uses dual cpus for some of the filters and functions is irrelevant to whether it's threaded or not, it only EVER does 1 thing at once and you have to wait for it to finish whatever that thing might be before you can ask it to do something else. All a batch does is save you from opening files yourself, all actions do is play back what you want them to do.

If this was a threaded in some way you could batch process while working on another image or images, all with seperate filters and image manipulations happening simulataneously. You can't.

OS 8 and later 9 did benefit from this multi-threading. The control strip for one did a great job of running alongside any hardware-intensive application. Audio CDs played non-stop even when rendering movies with QuickTime Pro. I am sure you can come up with your own examples.

An Audio CD playing in the background uses ZERO cpu time, other than sending a request for the first track and telling the drive to start playing the CD. It doesn't tax the hardware in any slight way, all it's doing is pushing the output of the CD into the audio hardware and then it's coming out of the speakers. About the only thing I can think of where it DOES use CPU time but doesn't interfere too much with other applications is playing mp3s. playing them in QT is the exception to that, it's worthless for playing mp3s because even opening a folder with a few dozen files can interupt playback for a split second.

Sorry if this nit picky, it's just true. The only real world use of multithreading under OS 8 - 9.x that I'm aware of is copying files in the finder between multiple drives and folders.

[Sol]

Currently an Apple Display is the best value PC monitor on the market. The price-drops are necessary to off-set people's response to war: saving money. The cynic would say that Apple's products are toys for the rich but Mac users know that they are better value than any PC (generally Macs do have better re-sale value than same-generation Windows hardware).

If times were better for the computer industry we may have had the video iPod all ready. G4s would probably be running at the same speed but MHz never worried me.

[Rocketman]

Originally posted by Sol
Your idea of a multi-threaded Photoshop sounds similar to batch-actions (you know, converting a folder of images to another format or something like that).

OS 8 and later 9 did benefit from this multi-threading. The control strip for one did a great job of running alongside any hardware-intensive application. Audio CDs played non-stop even when rendering movies with QuickTime Pro. I am sure you can come up with your own examples.

For its day, and for any single CPU G3 Mac, OS 8 and especially OS 9 were very good OSs. The move to a UNIX-based OS was not what I expected it to be but let me just say that for a multi-processor system this is the only OS I would want to use.

If Apple read this and other related messages and simply improved multi-tasking alot it would make a huge difference in day to day operations by users.

Rocketman

[MacBandit]

I think that a big part of the reason that Apple separates there consumer and pro machines by cpu and processor speed is. A) The price of the processor. Face it to sell these things at a good margin you can only cut the cost in so many places and still turn out a good product with name bran hardware. B) Supply of cpus. I think this is the main one. Home many G4s @ 1.42 do you think there are available? This has been true for a very long time for Apple. Because they are smaller they just can't drive the demand for enough of the higher processors to get Motorola to figure out how to crank out more of them. The faster the processor the lower the production yields.

[jettredmont]

Originally posted by bentmywookie
Ok, for those arguing about the single processor vs. dual processor issue, all I want to say is that an OS (at its lowest level) needs to do resource allocation/management. It's the interface between the hardware and the software.


True.


So, to characterize the scenario, a piece of software comes along and tells the OS, "hey I need to cook this meal, here's the recipe, now off ya go!" And the processor looks and sees, "well I have two ovens, 3 bowls, 5 fridges, etc." and basically it uses whatever it has to get the job done. Ok, enough with this analogy (for now).


Bad analogy, because the recipe most software follows is "do this, THEN do this, THEN do this", not "do this and this and this in any order". The string of things that have to be done in order is called a thread of execution. MOST, not all, software is written with a single thread of execution for the "main task" of the software. Like baking bread: you have to mix the ingredients, then knead it, then let it rise (iterate), then bake. If you bake it first, you end up with fried eggs and singed flour. No matter how many cooks you have in the kitchen making one loaf of bread, the process can not be sped up because the labor can not be divided up and made parallel instead of in series.

That is not to say that most software is single-threaded. In fact, I would be surprised if any OS X or in fact any post-Win95 software that does any amount of work whatsoever were single threaded, simply because then the app would be unresponsive to both the user and the OS while it did its work. Hence, apps that "do work" tend to do their "work" in a background thread while the "main" thread continues to respond to user and OS messages. It is possible (although not common), in this common case, for the message-handling thread to execute on one processor while the "work" thread executes on the other processor, leaving the work thread a bit more "room" on its CPU. It is not common for this to happen as the OS will preferentially place threads of a single application on the same processor (unless they both use full timeslices and the other CPU is relatively unused) as threads within an app are far more likely to share data between each other (and to require cross-thread signalling and mutexes) than threads in different apps, and these things are more efficiently done when the threads are on the same physical processor.

What multiple processors buys you, in a single-threaded-application world, is the ability to run two separate applications on two separate processors. If you had one app that is running full-bore and consuming massive swaths of CPU time, and a dozen other apps, the kernel will schedule the boorish app on one CPU and the other apps on the other CPU. Likewise, if you have two CPU-hog apps running, each wil be scheduled on its own CPU and the rest of the processes in the system will divide up amongst the two CPUs on top of the two hogs.

Third scenario, an app is written to divide its "main work" between two threads. The OS initially schedules both threads on one CPU, then quickly sees that they are both "CPU Hog" threads and pushes one over to the other CPU next timeslice. Thus, the OS quite efficiently can use its two processors even though the app itself wasn't "strictly" written to an SMP API.

Fourth and final scenario, like the third, except the app uses the SMP API. When the second "work" thread is spawned, the application tells the kernel that it will be a CPU hog and should be placed on a different CPU from the first if possible. This removes the couple of timeslices where both threads share the same CPU before the kernel sorts out that they need to be separated.

Fifth, an improvement on the fourth, the app asks the Kernel if it has a low-usage CPU available, and only if there is a second CPU available for processing does it spawn a second thread for processing. This is a bit more difficult to program, but often significantly improves performance on the single-CPU case while not affecting multiple-CPU performance.

So, going from one single-threaded app to multiple single threaded apps to a multiple-threaded app to a multiple-threaded app that is SMP-aware, you get increasingly efficient use of the second processor.

OS X does a pretty good job of dividing up the work here, and of course you'll never see a "single" app running on OS X because OS X itself consists of multiple background processes. But, you won't see your main app run significantly faster on a dual-proc machine than on a single-proc machine if it is single-threaded. The only advantage the dual-proc machine has is that OS X's background processes are shunted over to one CPU and thus your main app has the whole CPU to itself (but, of course, not the whole System Bus or Memory Bus or disk, etc ...)

Now, a word about benchmarks: Benchmarks tend to be sequential, not parallel. By this, I mean, one task is executed while the machine is doing nothing else, then another task is executed, then another, and so on. If the apps handling the tasks are single-threaded, you have the first scenario above (one CPU hog thread on one CPU and the OS threads on the other), which shows very little advantage to having multiple CPUs. As a result of the fact that most of the time the popular applications out there have a single "work" thread, this skews benchmarks heavily in favor of single-CPU systems. This type of benchmarking is realistic for single-task servers whose only interaction with users is how fast it got its job done. However, "real life" with a desktop or workstation computer is generally not like this. People multi-task, and get annoyed when their computers do not. Yes, they want to be able to browse the web and print that report out and listen to iTunes while their CD is burning. Dual (or more) processors allows this to happen, and that is never shown in benchmarks.

[jettredmont]

Originally posted by iAndy
Sorry this doesn't prove that PC are difficult to upgrade, only that you might not be too hot recognizing one end of a screwdriver from another ;)


Quite true, regarding my failed attempt at upgrading (nothing to do with mechanical incompetence, just with not contacting NEC to see if their MB was capable of upgrading before going out and buying the upgrade card based on the chipset and BIOS information). As is always true, anectdotal evidence proves nothing, just adds a story to tell.

However, my decision not to (generally) upgrade my CPU is purely financial, and not at all related to some mechanical ineptitude (I do quite fine swapping in and out PC parts, thank you!). Noting that I have the willpower to hold off on a new computer for a few years, not requiring the latest/greatest every 6 months, I find I more often than not have a use for the old box, or can find someone to buy it or at the very least a school that needs it. Thus, my choice is, spend $500 (assuming a new MB/innards) to end up with one faster CPU in a two-year-old system with a small HD and rickety video card, or spend $2000 to end up with a faster overall system AND add a new system to my server farm AND bump one of the oldies out for sale or donation. One computer or two? I find that the monetary difference is outweighed by the difference in utility, and once every couple of years I can afford it.


That is not necessarily the case. There are a number of companies out there making a very healthy living from cpu upgrade cards.


Ah, yes, the upgrade cards. I had looked into those and dismissed them way back when because they were reportedly highly unstable and had physical issues in my case (ended up too tall and blocked the video card slot). But, yes, I imagine they've gotten better since then (and i can't remember the last time I heard about someone with an upgrade card having a flaky system because of it).

Note the following review of a recent add-in card at Tom's Hardware:

http://www6.tomshardware.com/cpu/20030107/index.html

ASSUMING your BIOS will support a newer processor, the PowerLeap solution can work.

Upgrading to a 1.4GHz P3 cost them $250 (assuming tax was the reason ... they later noted that the upgrade package itself was just under $160), and certainly out-performed the old 866MHz P3 they had, but didn't come anywhere near to what you would see with a real 1.4GHz P3 system (which should have been on-par with if not slightly better than the 1.7GHz P4 results they showed).

So, yes, you can use a card and "sorta" get an upgraded CPU (living with the deficiencies of the rest of your system and the added performance hit of the daughtercard). But don't expect (like most people seem to expect when they start talking about upgrading their CPU) top-of-the-line performance anywhere near that of a whole new box.

This is very analogous to the Mac situation: you can add a CPU-on-a-card upgrade (and people do), but you shouldn't go in expecting 1.25GHz performance out of your old Cube.


Boy are you digging a hole for yourself here ! ;) Check out www.Powerleap.com for info on their latest 1.4GHz upgrade offering @ $160 - enough said !


I stand corrected. Do note that that is a Celeron, not a P3 (you have to hit the "Add to Order" button and read the fine print there ... Celerons are recommended for BX-based motherboards anyways, which is the most common situation), but, yes, a fairly good upgrade for the money.

Now, back to the scenario I was talking about (because it's the one that would conceivably give you a "fully-functioning" high-end CPU and because the post I was replying to specifically talked about replacing their ZIF processor with another processor): I would love to see you buy a 3.04GHz P4 and fit it in a motherboard which had housed a 2.0GHz P4. Doesn't work, because Intel frequently changes the socket form factor to prevent precisely this. While you could possibly add a little bit of speed to that 2.0GHz motherboard (depending on which line ot 2.0GHz P4s you got, the "2.0" or the "2.0A"), a direct chip-for-chip replacement is generally a one-time upgrade, and a minor one at that. Unless you upgrade the motherboard, etc.

[jettredmont]

Originally posted by eric_n_dfw
Small threads are used all over the place. You seem to think that coding them is in some way difficult. It is not. You code your objects to be thread safe (fairly trivial if you know what you are doing), you instantiate one and tell it to run. Sometimes you create a bunch and (in Java terminology) put them in a thread pool, ordering all of them to fire off at the same time.


Easy. And dangerous. Win NT dies at 500 threads on the system. Note that a thread lasts a little longer than the work you are making it do, and that creating/destrying threads is fairly expensive.


I've never looked at Mozilla's code but if you think about it, threads would be a obvious way to pull in images for an html page. You fire off 1 for each image and the independantly load the graphics without the main thread worrying about them.


Well:

1) I can assure you that the "main thread" of the application is not the same as the socket-communications thread. In any app. Period. Socket communications requires its own thread per connection, or a lot of pretty fancy programming.

2) HTTP 1.1 uses "persistent connections" which allow multiple items (the main page and its images, for instance) to be pulled from a single TCP/IP connection. This is important because TCP/IP connections are fairly expensive to make. I suspect that Mozilla "always" uses the persistent connection speed-up instead of using multiple threads (which would be significantly slower on a narrow-band connection than a single-thread download), jsut so it doesn't have to worry about multiple routes of communications.


I'd presume that Final Cut Pro does similar type things - firing off a pair of threads to render one frame each.


Way too granular. But, yes, having one thread handle, say, three keyframes of video, and another thread handling the next three keyframes, should be fairly efficient. Except - oops - the key bottleneck there is likely to be the disk and system bus bandwidth, not processing speed.


Don't believe me about the browser point? Try OmniWeb and open the inspector that shows the page rendering status (I can't remeber it's actual window name) - each element of the page shows up and reports it progress - you can kill any one element. This is multithreading.


Well, it's one way of doing it. I'm not going to say it's the "right" way because there are trade-offs involved (the hit on your internet connection and the hit on the server included). But, yes, if the "base" case for optimization included a long network lag to the server, plenty of processing power on the server, and a fat net connection between web browser and server, then this is a good approach.

Back to the original point (I think): Mozilla is an example of an app that uses one thread to do it's "work". IMHO, Mozilla is really a poor example all around regarding threading. There are times when Mozilla becomes completely unresponsive on Windows XP while downloading HTTP headers ... very poor implementations in there somewhere (and, yes, I should go fix it if I know so much, but damn ... "in there somewhere" is a far cry from knowing where to look in that monster!)

[jettredmont]

Originally posted by jettredmont

1) I can assure you that the "main thread" of the application is not the same as the socket-communications thread. In any app. Period. Socket communications requires its own thread per connection, or a lot of pretty fancy programming.


Oops, HTTP is one of the times when one thread-per-socket is not required (because the client doesn't need to respond to the server sending data back ... so long as the machine TCP buffers don't overflow, which is distinctly possible if you aren't careful ...). But still, I wouldn't want to do socket communications in the main thread of a graphical application. Too messy to code and thus to maintain.

[ddtlm]

jettredmont:

1) I can assure you that the "main thread" of the application is not the same as the socket-communications thread. In any app. Period. Socket communications requires its own thread per connection, or a lot of pretty fancy programming.
You've severely over-asserted this claim. Threads are one approach, however threads are definately not needed for low-data-flow applications. In fact, I think I could construct a number of examples where threads are an inferior way to manage sockets.

That is not to say that most software is single-threaded. In fact, I would be surprised if any OS X or in fact any post-Win95 software that does any amount of work whatsoever were single threaded, simply because then the app would be unresponsive to both the user and the OS while it did its work. Hence, apps that "do work" tend to do their "work" in a background thread while the "main" thread continues to respond to user and OS messages.
I guess you didn't read through my posts at all where I pretty much said the same thing ;) , and then dismissed this as irrelevent in the single-CPU vs dual-CPU debate. I'd be amazed if the performance gained by checking for keystrokes or whatever is greater than the performance lost by the SMP-OS constantly swapping the worker thread from CPU to CPU (with the OS overhead and the extra cache misses this causes).

[Wash!!]

This conversation is too rich for my blood

[jettredmont]

Originally posted by ddtlm
jettredmont:


You've severely over-asserted this claim. Threads are one approach, however threads are definately not needed for low-data-flow applications. In fact, I think I could construct a number of examples where threads are an inferior way to manage sockets.


Correct. As I stated in my next reply.

But, to be clear, if you're not using a thread per socket connection, you have to be sure that:

1) You "get around" to each socket before the local (machine-defined) TCP buffers overflow. TCP buffer overflow is messy, and can cause system-wide misbehavior at least on Windows.

2) There are no assumed "response times" which conflict with your worst-case reading lag.

This is okay for HTTP connections generally, assuming you have a reasonable "round-robin" time

I guess you didn't read through my posts at all where I pretty much said the same thing ;) , and then dismissed this as irrelevent in the single-CPU vs dual-CPU debate.


Guilty. Replied before reading the whole thread (I mean, it's 14 pages long!) Sorry for repeating!

I'd be amazed if the performance gained by checking for keystrokes or whatever is greater than the performance lost by the SMP-OS constantly swapping the worker thread from CPU to CPU (with the OS overhead and the extra cache misses this causes).

In theory, the kernel shouldn't be swapping any process back and forth between processors for just that reason (cache misses). Are you sure that OS X is doing this? If you can document it, I'd suggest shooting a bug report off to Apple, 'cause it is a bug. The kernel should switch a thread from one processor to another only when there is a "significant" advantage to doing so. "Significant" is fuzzy, yes, but it implies that the switch will last more than one or two time slices!

[ddtlm]

jettredmont:

In theory, the kernel shouldn't be swapping any process back and forth between processors for just that reason (cache misses). Are you sure that OS X is doing this?
I'm pretty sure, based mainly on the CPU monitor usually showing each CPU 50% used when one worker thread is going full out. I saw this in the Linux 2.2.x days, but the 2.4.x kernels fixed it. I hope OSX fixes it some day as well.

[macphisto]

Look how far off the topic of this thread has gotten. Kinda funny. It would be cool see a visual representation (in tree form, of course) of the many tangents people go off on in a discussion like this.

:)

BTW, where are the new iMacs!!?!?!!! :confused:

[ddtlm]

macphisto:

BTW, where are the new iMacs!!?!?!!!
Apple can't just go and do new product releases every day. Give them a few weeks.

More on topic: damn it is hard to pass up on the top-end PMac and 23" screen. An actual decent deal. I could get the Sony version of the screen but it costs $800 more (or so) and it's only real benefit is a second video input.