PowerMac G5 2.5GHz Benchmarks

[MacRumors]

One user provides some of the first benchmarks from Apple's newest top end PowerMac - the dual 2.5GHz model. Apple introduced the dual 2.5GHz PowerMac G5 in early June (http://www.macrumors.com/pages/2004/06/20040609083915.shtml) but the top end model is not due to ship until late July.

The following Cinebench benchmarks were run on a Dual 2.5GHz PowerMac with the new GeForce 6800 Ultra video card (connected to two 30" Apple LCDs).

For reference, Dual 2.0GHz PowerMac/Radeon 9600 Pro benchmarks from Xlr8yourmac's overview (http://www.xlr8yourmac.com/G5/Powermac_Dual_g5.html#storytop) are provided in parenthesis.

CINEBENCH 2003 v1
****************************************************
Processor : PowerMac G5
MHz : 2500
Number of CPUs : 2
Operating System : OS X 10.3.4
Graphics Card : GF 6800 Ultra DDL
Resolution : 2560x1600
Color Depth : millions
****************************************************
Rendering (Single CPU): 356 CB-CPU (Dual 2.0 = 219)
Rendering (Multiple CPU): 633 CB-CPU (Dual 2.0 = 401)

Multiprocessor Speedup: 1.78 (Dual 2.0 = 1.83)

Shading (CINEMA 4D): 335 CB-GFX (Dual 2.0 = 266)
Shading (OpenGL Software Lighting) : 995 CB-GFX (Dual 2.0 = 661)
Shading (OpenGL Hardware Lighting) : 1794 CB-GFX (Dual 2.0 = 1168)

OpenGL Speedup: 5.36 (Dual 2.0 = 4.40)

Aside from a 25% increase in processor speed, the new machine also houses a faster front-side bus (1.25GHz vs 1.0GHz) and improved video card.

[SFNE Freak]

Looks like it performs noticably better!

[Grimace]

yowsers - I want one now - and a 30" monitor!

[Don't panic]

not bad: 50% increase in performance!

[arn]

Usual caveats.... benchmarks measure very specific things, etc...

however... the rendering benchmarks are proportionally higher than expected from just the processor jump (25% processor speed bump vs 60% improvement in the single cpu rendering)

[swingerofbirch]

wouldn't running the two 30" monitors slow down the computer, if only marginally?

[legion]

I'd attributed most of the overall performance increases to the graphics card as cinebench offloads a lot of these tasks to the GPU and the GeForce6800 DDL is quite a bit faster than the previous.

As Arn's disclaimer before, cinebench isn't a great test to see processor performance as much as GPU+SystemBus+CPU.

[toughboy]

did anyone realise that the systems have different graphic cards??

for that reason the benchmarks may not show a real difference..

[clr900]

How did someone already get a dual 2.5 with a 6800 ultra, I thought they are not shipping until August? I plan on getting one too but didn't think that they are shipping right now.

[legion]

wouldn't running the two 30" monitors slow down the computer, if only marginally?

No, all the work is done by the graphics card (plus, dual link cards can easily go higher than the resolution of the 30" monitors, so it still isn't at a taxable level)

[benoda]

I wish the benchmark comparisons where more relevant. It's hard to discern what advantage the new video card gives over what the faster processor/bus provides.

I suppose all will be answered once they ship :)

[legion]

How did someone already get a dual 2.5 with a 6800 ultra, I thought they are not shipping until August? I plan on getting one too but didn't think that they are shipping right now.

WWDC setup (note the monitors too!)

[arn]

How did someone already get a dual 2.5 with a 6800 ultra, I thought they are not shipping until August? I plan on getting one too but didn't think that they are shipping right now.

WWDC

arn

[themadchemist]

Despite the obvious fact that there are multiple variables here, it's nice to see how the high end package available now does against its predecessor. Of course, cinebench only tests a few things, but it seems that there is significant improvement, largely because that sucktastic graphics card is gone.

[clr900]

Oh alright, it is nice to have some insider info!

[Sir_Giggles]

Now that Mac OSX will use Core Image and Core Video to offload some processing to the GPU, this throws a wrench into some of the benchmarks for rendering. As image-manipulation apps and 3d programs are sure to benefit greatly from GPU rendering, are we to see Cinebench being updated to allow for these new benchmarks?

[Santaduck]

if the powers that be that leaked the 2.5 dp cinebench results are reading this, could you try an Unreal Tournament 2004 benchmark? Both botmatch Asebestos, and flyby Asbestos on "max" detail settings (so we can compare to the xlr8yourmac database (http://forums.xlr8yourmac.com/fpsdb/fpsdb.lasso). Botmatch tends to be CPU-limited, and the flyby gfx-limited.

Benchmarker for UT2k4 is available here (http://www.santaduck.com/ut2k4/Santaduck_Toolpak_pub.sit).

[realityisterror]

no fair!!!
i want one of those baby's.
how come it only recognizes one monitor?

reality

[msconvert]

Now that Mac OSX will use Core Image and Core Video to offload some processing to the GPU, this throws a wrench into some of the benchmarks for rendering. As image-manipulation apps and 3d programs are sure to benefit greatly from GPU rendering, are we to see Cinebench being updated to allow for these new benchmarks?

Does anyone else remember last years WWDC? One of the graphics programs was specifically reprogramed to offload the GPU onto the G5 because it was actually faster. (The 3d world generator). Now the card is faster for floating point calculations? What gives? Is the card that much better - Stevo did say in the current WWDC keynote that GPUs are out pacing moores law so maybe I answered my own question. Or they are prepairing for the future if it isn't already here....

Thoughts?

[Sabbath]

Looks like a nice little boost, hopefully we will keep seeing decent boosts in the G5 scores as the FSB will keep going up with the chip and memory and GPUs keep getting faster all the time.

I know the nvidia is a serious step up from the 9600 maybe it would be better compare against the 9800, but it still is a decent improvement.

[kirk26]

Sweet, now you babies can stop asking if this is enough to have 2 internet pages open while working Photoshop!

[slughead]

Looks like it performs noticably better!

Looks like the 6800 is a 6800. Both computers need need the same video card to compare the processors in most of these tests.

For reference, Dual 2.0GHz PowerMac/Radeon 9600 Pro benchmarks from Xlr8yourmac's overview are provided in parenthesis.

I'm waiting for a real comparison, as all half of the comparisons mean nothing.

Move along, nothing to see here.

[PlaceofDis]

wow now only if i could afford a setup like that

[jadedchameleon]

Well, the rendering tests are really CPU tests. The shading tests are more graphics card tests. Rendering scores among the same setup with different video cards tend to be the same. I would say the speedup seems quite good from 2.0 to 2.5.

[CholEoptera36]

Very nice increase in performance from upgrading. I like the numbers. :) Anybody got any other figures or markings to add to these?

[Steven1621]

does anyone know if the new displays are in apples stores yet?

[pounce]

i know precisely what these numbers mean.



they mean i'm coming unglued with anticipation of getting my dual 2.5 g5, 2 gigs ram, a terrabyte of hd space, etc. it's gonna be a sweet ride for me. as an audio engineer, the track counts, plug ins, and higher sample rates will take advantage of the extra power these new macs have. i'll still keep the dual g4, but it looks like the g5 will be so big an improvement that i won't look twice at the g4 anymore. wow.

[ddtlm]

Those look like video card benchies, not CPU benchies, and its no suprise that the 6800 destroys all the came before it. CPU/FSB/RAM benchmarks would be far more interesting.

[ziwi]

Impressive....most impressive.

[[G5]Hydra]

Sorry, but they are wrong. Nobody would be happier than I to see the new G5 kick some serious butt, but there are a few things that need to be addressed. Cinebench has two flavors for OS X. One is a G5 targeted beta which is much faster on a G5. The older version, the one linked to in the XLR8 page is obviously the older one which I get similar results to on my own dual 2.0. Here is the correct result for the G5 version on a dual 2.0 (Rev A.):

CINEBENCH 2003 v1
****************************************************

Tester :

Processor : G5x2
MHz : 2000
Number of CPUs : 2
Operating System : 10.3.4

Graphics Card : 9800 OEM
Resolution : 1680x1050
Color Depth : <fill this out>

****************************************************

Rendering (Single CPU): 288 CB-CPU
Rendering (Multiple CPU): 503 CB-CPU

Multiprocessor Speedup: 1.75

Shading (CINEMA 4D) : 278 CB-GFX
Shading (OpenGL Software Lighting) : 783 CB-GFX
Shading (OpenGL Hardware Lighting) : 1494 CB-GFX

OpenGL Speedup: 5.37

****************************************************

I was lazy and didn't quit any apps and just dl'd the app from Maxon's site and ran the bench. If I rebooted and quit all other apps I could have gotten a bit better performance. The new dual 2.5 is fast but it is closer to the 25% clock speed increase, I think my benches are pretty close in fact to this number. It won't be anywhere near 50% faster. I just registered because I read MacRumors often and didn't want that error on the front page.

-Jerry C.

[michaelb]

I'd attributed most of the overall performance increases to the graphics card as cinebench offloads a lot of these tasks to the GPU and the GeForce6800 DDL is quite a bit faster than the previous.

Since when is raytracing (rendering a final scene) offloaded to a graphics card at all...?

(If raytracing could be done in the video GPU, imagine the beautiful artistry that would appear on screen when playing Doom 3!)

Cinebench's first two tests (single-processor and multiprocessor rendering) are about as CPU intensive as a program gets - raytracing involves millions of high precision floating point calculations, none of which can be offloaded to a graphics card.

A good graphics card influences fast screen updates while modeling yes, but final rendering no - a hunking great processor or render farm is what you need there.

[WorldMage]

Does anyone else remember last years WWDC? One of the graphics programs was specifically reprogramed to offload the GPU onto the G5 because it was actually faster. (The 3d world generator). Now the card is faster for floating point calculations? What gives? Is the card that much better - Stevo did say in the current WWDC keynote that GPUs are out pacing moores law so maybe I answered my own question. Or they are prepairing for the future if it isn't already here....

Thoughts?

This depends a lot on what the calculation is.
So General purpose CPU's like the G5 tend to be good at 'complex'
algorithms, things that have a lot of logic mixed with the math.
But graphics cards are _amazingly_ fast at doing the same
'simple' operation to many many pixels. They have great access
to memory and generaly have 8-16 little processors that all work
at the same time. But the processors aren't very flexible so if the
algorithm needs 'logic' it tends to get complex/slow/impossible to
do on the GPU.

Core Image is looking to move these simple things to the GPU.
Typically the bottle neck here is bandwidth to the card, so the
image spends more time going from main memory to the card
and back again (assuming it's needed in main memory at the
end of everything) than actually being processed by the card.
I think the card interfaces are speeding up so this isn't as big
a deal.

[DavidCar]

I think the card interfaces are speeding up so this isn't as big
a deal.

I wonder how long it will take for the G5 to get a PCI-Express interface for its graphics cards. Six months?

[nallu]

i would really like to know how this one compares to the bare-feats g5, opteron and xeon benchmarks.

their values are not in secs but in cf ratings can any one help


thank you

[rendezvouscp]

That's a lot more than I thought they would increase. Wow.
–Chase

[Fitzcaraldo]

Any word on noise levels with the 6800 Ultra inside?

[SwitchingSoon]

Sweet mother of god. I can't wait.

I wonder how my baby will play on my Max resolution of 1024x768 VGA gateway monitor! ooohh jeaaah....

*cries*

[michaelb]

I wonder how long it will take for the G5 to get a PCI-Express interface for its graphics cards. Six months?

Depends on demand I guess. PCI-Express is in Intel's latest chipsets mainly because they want to shake up the status quo and get everyone to upgrade, not I think because graphics cards actually need it.

My understanding was that for the immediate future, AGP was just fine. Can graphics card yet saturate the 8x AGP bandwidth, let alone the 16x PCI-Express?

Mainboard makers are actually having to include both PCI-X (for backward compatibility) and PCI-Express, making for larger boards and bigger tower cases. I'm not sure Apple should go that route yet.

Once more cards are available for PCI-Express, maybe then's the right time.

[~Shard~]

does anyone know if the new displays are in apples stores yet?

I thought they weren't shipping until July, with the 30" and corresponding video card not shipping until August - I thought that's what Jobs said in the Keynote...

[ITR 81]

I should get my 2.5 G5 on the first week of Aug. if not slightly earlier.

I don't have the top graphics card but the std one is ok for my quick image editing for my biz.

If I buy one for home it will come with a 30in LCD and the 6800.
Maybe by Jan they will have the 3Ghz G5.

[macridah]

good upgrade. i see not only the increase in cpu helps boost performance, but also the faster bus speed. bus speed is very important.

[legion]

Mainboard makers are actually having to include both PCI-X (for backward compatibility) and PCI-Express, making for larger boards and bigger tower cases. I'm not sure Apple should go that route yet.

PCI-X and PCI-Express have nothing to do with each other. They have completely different purposes. For now, if a mobo maker had both AGP and PCI-Express, then you'd have a valid point. The PCI-X on these systems isn't because of "backward compatability."

[DMann]

Well, the rendering tests are really CPU tests. The shading tests are more graphics card tests. Rendering scores among the same setup with different video cards tend to be the same. I would say the speedup seems quite good from 2.0 to 2.5.

Percentage exceeded what most expected.......

[DMann]

good upgrade. i see not only the increase in cpu helps boost performance, but also the faster bus speed. bus speed is very important.


And literally a quantum leap from the bottlenecked
G4 systems - remember the sad days of the 100MHz bus?

[DMann]

Any word on noise levels with the 6800 Ultra inside?

I can only imagine that it is quieter than the ATI 9800, and the
entire line of ATI cards---noisy little fans....

[DMann]

I wonder how long it will take for the G5 to get a PCI-Express interface for its graphics cards. Six months?

Hopefully by the next generation, likely to be
introduced by January of next year.

[legion]

I can only imagine that it is quieter than the ATI 9800, and the
entire line of ATI cards---noisy little fans....
There's a fan on the Ultra DDL (hence part of the reason it takes up one of the PCI-X slots.)

[DMann]

does anyone know if the new displays are in apples stores yet?

Probably in stores by mid-July....

[DMann]

How did someone already get a dual 2.5 with a 6800 ultra, I thought they are not shipping until August? I plan on getting one too but didn't think that they are shipping right now.

August - mid August is expected.

[dizastor]

Since I didn't win the $290 million lottery tonigh, I guess I won't be getting my dual 30" screens after all.

[DMann]

Since I didn't win the $290 million lottery tonigh, I guess I won't be getting my dual 30" screens after all.

$290 million in NJ, 10 rounds, not even close. Perhaps
I'll stop playing and put the money into a dual 30" screen
fund :rolleyes:

[Kingsfoil13]

I do most of my work in Cinema4D, so the Cinebench scores always seem particularly relevant to me. (before someone asks: Maxon makes Cinebench, and it is basically just a stripped-down version of their 3d modeling app, Cinema4D.)

I'm running a stock rev A 2 x 2Ghz G5 (with an ATI 9600 Pro) and 2GB of crucial RAM added (for what that's worth) for a total of 2.5GB. The monitor in question is a NEC-Mitsubishi 1760vm, 1280x1024, running through an ADC-DVI adapter and then a Belkin DVI cable direct to the monitor. Here are my specs:

CINEBENCH 2003 v1
****************************************************

Tester : Kingsfoil13

Processor : Dual 2.0 G5 rev A
MHz : 2 x 2.0 Ghz
Number of CPUs : 2
Operating System : 10.3.4

Graphics Card : Ati 9600 Pro
Resolution : <fill this out>
Color Depth : <fill this out>

****************************************************

Rendering (Single CPU): 287 CB-CPU
Rendering (Multiple CPU): 498 CB-CPU

Multiprocessor Speedup: 1.74

Shading (CINEMA 4D) : 276 CB-GFX
Shading (OpenGL Software Lighting) : 749 CB-GFX
Shading (OpenGL Hardware Lighting) : 1347 CB-GFX

OpenGL Speedup: 4.87

****************************************************

As you can see I am not in the range of the 2 x 2.5 with the Ultra, but neither am I hopelessly far out. I think I'll save my pennies and wait for the G6s.

Since most of what I care about is final-render time, who can tell me what hardware upgrades I should be concentrating on?

And on another note, how much of all of this is Tiger going to change by offloading everything and its mother to the GPU? I'm thinking I'll wait until I have Tiger, and then check the video card scene and upgrade to whatever helps photoshop / cinema4d work the most. But any advice would be appreciated.

[daveg5]

the benchmarks are whacked unless you have the same memory, hard drive and video card in both.

[The Bartender]

does anyone know if the new displays are in apples stores yet?

They arent... didnt Jobs comment on when they were due?

[Booga]

I'm glad others have corrected the original data, but it was really obvious that those figures were wrong. You can't increase the clockspeed by 25%, change nothing else, and get >50% speed increase. In fact, you can usually be confident that percentage performance won't even keep up with percentage clockspeed increase. And you should never publish uncontrolled benchmarks; as in this case, they could use different versions of the benchmarks, or there could be a hundred other variables to compensate for.

[dahacouk]

I'm interested in how these test match my current machine (Dual G4 1.25) so I quote:

http://www.xlr8yourmac.com/G5/Powermac_Dual_g5.html

Cinebench 2003 (v1) results: (NOTE: Scores in Italics are from a MDD Dual G4/1.25GHz w/9800 Pro card running 10.2.6)

****************************************************

Tester : EMM

Processor : PowerMac G5
MHz : 2000
Number of CPUs : 2
Operating System : 10.2.7

Graphics Card : Radeon9600Pro
Resolution : 1024 x 768
Color Depth : Millions

****************************************************

Rendering (Single CPU): 219 CB-CPU (Dual G4 1.25GHz/9800 = 120)
Rendering (Multiple CPU): 401 CB-CPU (Dual G4 1.25GHz/9800 = 220)

Multiprocessor Speedup: 1.83 (Dual G4 1.25GHz/9800 = 1.83)

Shading (CINEMA 4D) : 266 CB-GFX (Dual G4 1.25GHz/9800 = 149)
Shading (OpenGL Software Lighting) : 661 CB-GFX (Dual G4 1.25GHz/9800 = 411)
Shading (OpenGL Hardware Lighting) : 1168 CB-GFX (Dual G4 1.25GHz/9800 = 875)

OpenGL Speedup: 4.40 (Dual G4 1.25GHz/9800 = 5.87)

****************************************************

Cheers Daniel

[mac15]

Very impressive, I think Pro's are going to absolutely love a Dual 2.5/30" Combo!

[XDream]

these benchmarks justify my decision - i've already ordered a 2.5 G5 a few days ago :)

[Veldek]

I can only imagine that it is quieter than the ATI 9800, and the
entire line of ATI cards---noisy little fans....

It seems the 6800 is noticably louder than the 9800XT. Look here:

http://www.hardmac.com/niouzcontenu.php?date=2004-07-01#2414

[rdowns]

yowsers - I want one now - and a 30" monitor!

Just one monitor? Geesh, what a lightweight. :D

[rdowns]

How did someone already get a dual 2.5 with a 6800 ultra, I thought they are not shipping until August? I plan on getting one too but didn't think that they are shipping right now.

The tests were run by a WWDC attendee. Available there, not shipping yet.

[gekko513]

Depends on demand I guess. PCI-Express is in Intel's latest chipsets mainly because they want to shake up the status quo and get everyone to upgrade, not I think because graphics cards actually need it.

My understanding was that for the immediate future, AGP was just fine. Can graphics card yet saturate the 8x AGP bandwidth, let alone the 16x PCI-Express?

Mainboard makers are actually having to include both PCI-X (for backward compatibility) and PCI-Express, making for larger boards and bigger tower cases. I'm not sure Apple should go that route yet.

Once more cards are available for PCI-Express, maybe then's the right time.
And besides ... PCI Express is currently slower than AGP 8x for most things. It's better to wait until the technology matures. PCI Express vs AGP (http://www.anandtech.com/textdoc.html?i=2091&p=12)

[blutfink]

Even though xlr8yourmac's benchmarks are outdated and thus do not provide a valid comparison (as pointed out by several posts above), the rumor of the 2.5 GHz G5s being 60% faster than their predecessors will spread anyway... :rolleyes:

[LiciousTheDiva]

Shouldn't it read...

"One user provides some of the first benchmarks from nVidia's newest GeForce 6800 Ultra DDL card for PowerMacs." ?????????[I]

[eric67]

I do most of my work in Cinema4D, so the Cinebench scores always seem particularly relevant to me. (before someone asks: Maxon makes Cinebench, and it is basically just a stripped-down version of their 3d modeling app, Cinema4D.)

I'm running a stock rev A 2 x 2Ghz G5 (with an ATI 9600 Pro) and 2GB of crucial RAM added (for what that's worth) for a total of 2.5GB. The monitor in question is a NEC-Mitsubishi 1760vm, 1280x1024, running through an ADC-DVI adapter and then a Belkin DVI cable direct to the monitor.


Guys stop getting excited so much!!! Cinebench values published here by macrumors were obtained on a NON G5-optimized CineBench!!!
have alook here to see the difference on the results, when using a dual 2GHz (scroll down bottom of the page):
http://www.hardmac.com/niouzcontenu.php?date=2003-10-16#6891

I do not imply that the Dual 2.5GHz is not blazzing fast, it is just that those numbers do not really speak properly while comparing 2 G5...

[fixyourthinking]

Now that Mac OSX will use Core Image and Core Video to offload some processing to the GPU, this throws a wrench into some of the benchmarks for rendering. As image-manipulation apps and 3d programs are sure to benefit greatly from GPU rendering, are we to see Cinebench being updated to allow for these new benchmarks?

I agree completely.

What needs to be tested is DV file conversion/iTunes Mp3 conversion/and photoshop render - via version 7 that wouldn't be affected by the coreimage API.

XLR8yourMac usually provides these tests.

Checking Unreal tournament framerates would also help.

Running Xbench may not even work until the program were specifically set for these new CPUs.

It's still a VERY interesting view of the new PowerMacs - indicating to me that 3.0Ghz isn't even necessary to ship right now. Nothing could take advantage of the faster than 2.5Ghz/Ultra setup right now.

[shawnce]

I note some interest in OpenGL capabilities of mac systems, etc. In that regard I made the following post (http://discussions.playnet.com/viewtopic.php?p=1195014#1195014) in another forum that contains some interesting graphs comparing systems/adapter capabilities across various systems using the OpenMark benchmark tool. Not fully scientifically done but still interesting.

----

The following list are links to graphs (PDF format) for various systems and test configurations using the OpenMark OpenGL benchmarking tool. I will update this post as I complete tests (add in the links)...

Summary Results:

So the following graph is using numbers from the systems that I have the numbers for so far (posted to this topic or my own systems). I picked the 10k to 100k range of triangles because that appears to be a point past which the adapter and CPU/system factors both come into play. In other words the numbers appear to show more of a overall system limiting not so much an adapter limit. The first part of full range graph (0-100k) can give you the idea of how the adapters compare with each other before system limits are hit.

System Comparison at Simular Resolutions 10k-100k triangles (http://homepage.mac.com/shawnce/wwiionline/OpenMark/System_Compare.pdf) (updated 6/23)

System Comparison at Simular Resolutions 0-100k triangles (http://homepage.mac.com/shawnce/wwiionline/OpenMark/System_Compare_Full.pdf) (updated 6/23)

Detailed Results:

PowerMac Dual G5 2GHz (1st gen), 2GB RAM, ATI Radeon 9800 PRO 128MB, Mac OS X 10.3.4
(note this system has two monitors attached both at 1920x1200 by default, so only 64MB of VRAM is available per monior, also I had one CPU being consumed with a back ground task not related to OpenMark)

Surface Size: 1920x1200x32bpp (http://homepage.mac.com/shawnce/wwiionline/OpenMark/PM_D2_ATI9800PRO128_19x12.pdf)
Surface Size: 1600x1200x32bpp (http://homepage.mac.com/shawnce/wwiionline/OpenMark/PM_D2_ATI9800PRO128_16x12.pdf)
Surface Size: 1344x840x32bpp (http://homepage.mac.com/shawnce/wwiionline/OpenMark/PM_D2_ATI9800PRO128_13x8.pdf)
Surface Size: 1280x800x32bpp (http://homepage.mac.com/shawnce/wwiionline/OpenMark/PM_D2_ATI9800PRO128_12x8.pdf)
Surface Size Comparison (http://homepage.mac.com/shawnce/wwiionline/OpenMark/PM_D2_ATI9800PRO128_Compare.pdf)

PowerBook 15" G4 1.25GHz, 1GB RAM, ATI Mobility Radeon 9600 64MB, Mac OS X 10.3.4
Surface Size: 1280x854x32bpp (http://homepage.mac.com/shawnce/wwiionline/OpenMark/PB15_1_25_1280x854.pdf)
Surface Size: 1152x768x32bpp (http://homepage.mac.com/shawnce/wwiionline/OpenMark/PB15_1_25_1152x768.pdf)
Surface Size: 896x600x32bpp (http://homepage.mac.com/shawnce/wwiionline/OpenMark/PB15_1_25_896x600.pdf)
Surface Size Comparison (http://homepage.mac.com/shawnce/wwiionline/OpenMark/PB15_1_25_Compare.pdf)

PowerBook 12" G4 867MHz, 640MB RAM, nVidia GeForce4 420 Go 32MB, Mac OS X 10.3.4
...being worked on...


Current Raw Data (Excel workbook) (http://homepage.mac.com/shawnce/wwiionline/OpenMark/OpenMark_results.xls)

[shawnce]

I agree completely.

What needs to be tested is DV file conversion/iTunes Mp3 conversion/and photoshop render - via version 7 that wouldn't be affected by the coreimage API.

Note CoreImage is not involved at all in the tests being talked about, etc. In fact the systems at WWDC that have the 30" screens are not even running Tiger (Mac OS X 10.4).

[må¥å]

$290 million in NJ, 10 rounds, not even close. Perhaps
I'll stop playing and put the money into a dual 30" screen
fund :rolleyes:

By the time you save enough money in that Jar Apple whould have released a 44 inch screen. ;) :) :D

[shawnce]

PCI-X and PCI-Express have nothing to do with each other. They have completely different purposes. For now, if a mobo maker had both AGP and PCI-Express, then you'd have a valid point. The PCI-X on these systems isn't because of "backward compatability."

PCI, PCI-X and PCIe are related in many ways... so the above doesn't make much sense to me. The connectors (signaling used) for PCIe are not compatible with traditional PCI adapters and hence system boards ship with a mix of PCIe and PCI/-X for now. Down the road the goal of PCIe is to over take and replace legacy PCI slots not just AGP.

Note the following article... http://arstechnica.com/paedia/p/pci-express/pcie-1.html

[wdlove]

does anyone know if the new displays are in apples stores yet?

I hope that they will be on display soon. Plan to visit my local Apple Store next week. Even though they aren't shipping yet, if they had them at WWDC they should be at the stores.

[Duff-Man]

They arent... didnt Jobs comment on when they were due?Duff-Man says....this has probably been asked and answered in the thread about the displays but anyway....I just tried the online Apple Store and the new 20" display showed as 2-4 weeks. while the 30" came up as "august" - I would suspect they would get demo units into stores before they ship from the online store.....oh yeah!

[djbahdow01]

My new 20in monitor is shipping on or before 07/29/04. Which is one day before my Dual 2.5 G5. So i would hopefully get them around the same time.

[shyataroo]

ok big whoop the Dual 2.5Ghz G5 is slightly faster than its predicessor it could just be me but a dual 2.5 ghz is a total of 1ghz faster than the dual 2.0 and with a faster bus/GPU it should be somewhere between 50-75% faster but its not! and the liquid cooling is nifty to say but does it acutally use less power? it seems not I say that you people are blinded by Apple don't get me wrong a Dual 2.5ghz G5 with 8GB of ram and 1TB of HD space and a Geforce FX 6800 Ultra would be extremly nice but its still a underwelming improvment

[dragula53]

ok big whoop the Dual 2.5Ghz G5 is slightly faster than its predicessor it could just be me but a dual 2.5 ghz is a total of 1ghz faster than the dual 2.0 and with a faster bus/GPU it should be somewhere between 50-75% faster but its not! and the liquid cooling is nifty to say but does it acutally use less power? it seems not I say that you people are blinded by Apple don't get me wrong a Dual 2.5ghz G5 with 8GB of ram and 1TB of HD space and a Geforce FX 6800 Ultra would be extremly nice but its still a underwelming improvment

1) a second processor isn't 100% efficient.
2) the difference between 4ghz and 5ghz is 25%, not 50% or 75%

therefore...

3) I have no idea why you are tripping

[pounce]

"underwhelming improvement" is too strong a statement.

faster bus, faster and brand new processor, faster memory, faster newer better graphics cards, liquid cooling, etc.. lots of improvements of all kinds, and indications that it's one of the fastest machines out there. for similar money you can get a similar spec'd pc that is similar in speed, depending on which speed test you look at. but all the boring internet sissy fights about that ************ aside, this is a gorgeous computer. as a total package, and an off the shelf solution, between the os and the hardware this is a winner for sure.

[DMann]

It seems the 6800 is noticably louder than the 9800XT. Look here:

http://www.hardmac.com/niouzcontenu.php?date=2004-07-01#2414

Thank you for the info............... I'm surprised, and then, again,
I'm not surprised.

[DMann]

By the time you save enough money in that Jar Apple whould have released a 44 inch screen. ;) :) :D

Now that all depends on how many rounds of lotto, Pick 6,
Pick 5, Pick 4, and Pick 3 I play per day/week. With that
in mind, Apple will likely release a 37" by then :p

[GrizzlyHippo]

I'm in the market for a new G5, upgrading from a G4/400 AGP with a 1Ghz Powerlogix upgrade card.

The two things that have bugged me about my G4 for the last 5 years are:

1 - It won't go to sleep and wake up without crashing, upgrade card or not - How are the G5's with sleep?

2 - I'm really interested in a dual 2.5 G5, especially due to the water cooling - Does the water cooling make the machine quieter than the other two models, this is what I really want to know? Does anyone know?

Cheers,
Grizzly.

[DavidCar]

And besides ... PCI Express is currently slower than AGP 8x for most things. It's better to wait until the technology matures. PCI Express vs AGP (http://www.anandtech.com/textdoc.html?i=2091&p=12)

The above article notes that "The bottom line is that the real benefit [of PCI Express] will present itself to applications that require communication with the rest of the system, like video streaming and editing, or offloading some other type of work from the CPU onto the graphics card."

So probably CoreImage and CoreVideo would benefit the most from a move to PCI Express over AGP.

[DMann]

I'm in the market for a new G5, upgrading from a G4/400 AGP with a 1Ghz Powerlogix upgrade card.

The two things that have bugged me about my G4 for the last 5 years are:

1 - It won't go to sleep and wake up without crashing, upgrade card or not - How are the G5's with sleep?

2 - I'm really interested in a dual 2.5 G5, especially due to the water cooling - Does the water cooling make the machine quieter than the other two models, this is what I really want to know? Does anyone know?

Cheers,
Grizzly.

I have both the G4 Sawtooth 500Mhz, and the G5 2.0Ghz. The G5 is MUCH,
MUCH quieter than the G4. However, the new Invidia 6800 ULTRA graphics
card could add some noise to the comparatively quiet G5 with it's on-board
fan.

[fixyourthinking]

Note CoreImage is not involved at all in the tests being talked about, etc. In fact the systems at WWDC that have the 30" screens are not even running Tiger (Mac OS X 10.4).

But neither were the other tests I was talking about. Are you sure that the particular 30" display models were not running Tiger? The dualie units that "my person" that saw them; said it WAS on all the machines that had dual 30" displays hooked up to them.

*By the way - I do see 10.3.4 in the specs but it almost has to be a custom build of panther for the card and the display to work right. If not, these testes are really skewed due to no driver optimization.

And as i said, Tiger WAS on all dualie 30" setups there - di he boot into 10.3.4 on his own?

Again

[shawnce]

But neither were the other tests I was talking about. Are you sure that the particular 30" display models were not running Tiger? The dualie units that "my person" that saw them; said it WAS on all the machines that had dual 30" displays hooked up to them. I was at WWDC working on the systems that had the new displays connected. I didn't see Tiger installed when I was using them.

[Fitzcaraldo]

It seems the 6800 is noticably louder than the 9800XT. Look here:

http://www.hardmac.com/niouzcontenu.php?date=2004-07-01#2414

Thanks, I had seen the Hadmac article... Alas the Files I downloaded seem to be empty Audio files??? so none the wiser. :(

I assume MacBidouille will show another Verax conversion as per the ATI 9800. :eek:

Surely its time to start including the GPU in the water cooling arrangement for rev 3.

[virividox]

xbench is all well and good, but i want to see it in real world tests. i dont doubt that its better, but i wonder if its worth the price.

[shawnce]

I'm in the market for a new G5, upgrading from a G4/400 AGP with a 1Ghz Powerlogix upgrade card.

The two things that have bugged me about my G4 for the last 5 years are:

1 - It won't go to sleep and wake up without crashing, upgrade card or not - How are the G5's with sleep? Only had one sleep problem with my G5 and that was a result of flaky disk enclosure connected via Firewire. Disconnect it and no issues since. I sleep my system almost every night.

2 - I'm really interested in a dual 2.5 G5, especially due to the water cooling - Does the water cooling make the machine quieter than the other two models, this is what I really want to know? Does anyone know? I wasn't in a quiet room with the 2.5 G5 at WWDC so I cannot say directly... but they are quiet.

A dual G5 anything will way outpace the system you currently have (and bring it into truly being a Mac OS X supported system).

[phillymjs]

Does the water cooling make the machine quieter than the other two models, this is what I really want to know? Does anyone know?

I don't see how much quieter it could be-- right when they first came out, I had two 2.0 dualies to play with for a week while I set them up for a client. They were so dead quiet that when Energy Saver blanked the monitor, the only way you knew it was on without peering through the mesh at the the fans was by looking at the power light.

The only time the G5 was noisy was when I had it in target disk mode-- the fans were running flat out and you could feel the breeze from quite an impressive distance away.

~Philly

[m.r.m.]

just out of curiosity, how does one overclock a mac and how far can one go (without having to build in a prometeia mach 2 or something like that).

[dguisinger]

The above article notes that "The bottom line is that the real benefit [of PCI Express] will present itself to applications that require communication with the rest of the system, like video streaming and editing, or offloading some other type of work from the CPU onto the graphics card."

So probably CoreImage and CoreVideo would benefit the most from a move to PCI Express over AGP.


It is obvious you don't understand what PCI Express is. PCI Express is (AFAIK) a 4-wire RX/TX pair capable of pushing in serial more bandwidth than a normal PCI slot, and if I remember correctly it is a jump from 133MB/sec to 150MB. Someone can correct me on that if I'm wrong. That may not seem significant to you, however:

On standard PCI, ALL SLOTS SHARE the 133MB/sec bus! Thats right, you can only have so many signals on the board running in parallel, and PCI has nasty timing issues because all those slots must be kept properly sync'd with each other and properly terminated at the end of the bus.

In PCI Express, the slots are not tied together, the PCI Express chipset acts as a switch. If two cards are talking together instead of basically shutting down the whole PCI bus, the switch just relays the serial PCI data back and forth between the two pairs of RX/TX lines, effectively forming a private network between the two.

Similarly, the PCI Express spec specifies PCI-E x1, PCI-E x2, PCI-E x4, and PCI-E x16. PCI-E x1 is a small slot and is what you are referring to. The x1 is number of RX/TX pairs. PCI Express x16 was designed for graphics cards, and has 8x the bandwidth of AGP. That means the PCI Express x16 slot is longer and has 16 PCI-Express interfaces to the slot, its kind of like if you ever saw the quad ethernet cards that were out when gigabit was really expensive. You could setup 4 100BaseT connections to act like one and run in parallel, effectively giving you a 400BaseT link with inexpensive technology.

PCI Express is a far supperior technology to PCI, yet it remains software compatible. Infact, PCI Express's serial and switching technologies are not even visible from the OS side of things, all PCI devices are configured the exact same way.

And of course, may PCI Express chipsets have PCI Express to PCI bridges, built in, allowing both PCI, PCI Express, and PCI Express x16 slots on the same machine.

As it is right now, PCI is overwealmed with data. Look at the XServe. 3 Serial ATA controllers running at 150MB/sec. 2 Gigabit ethernet controllers. An optional Fibre Channel card...lets not forget USB2 and FireWire 800. All of this is funnelled into the SAME PCI BUS. There is no way you can throttle all these devices at the same time. Now, what happens with PCI Express? You are now limited to the bottle necks in other parts of the system....... memory throughput to the chipset.... and CPU speed / front side bus speed....... both of which the PPC G5 platform are doing very well at.

I for one, am under strong beleif, that a new XServe will be announced in the August / September time frame with PCI Express, as it can take advantage of the PCI Express throughput the most with all the integrated devices in the server.

[Mav451]

It is obvious you don't understand what PCI Express is. PCI Express is (AFAIK) a 4-wire RX/TX pair capable of pushing in serial more bandwidth than a normal PCI slot, and if I remember correctly it is a jump from 133MB/sec to 150MB. Someone can correct me on that if I'm wrong. That may not seem significant to you, however:

On standard PCI, ALL SLOTS SHARE the 133MB/sec bus! Thats right, you can only have so many signals on the board running in parallel, and PCI has nasty timing issues because all those slots must be kept properly sync'd with each other and properly terminated at the end of the bus.

In PCI Express, the slots are not tied together, the PCI Express chipset acts as a switch. If two cards are talking together instead of basically shutting down the whole PCI bus, the switch just relays the serial PCI data back and forth between the two pairs of RX/TX lines, effectively forming a private network between the two.

Similarly, the PCI Express spec specifies PCI-E x1, PCI-E x2, PCI-E x4, and PCI-E x16. PCI-E x1 is a small slot and is what you are referring to. The x1 is number of RX/TX pairs. PCI Express x16 was designed for graphics cards, and has 8x the bandwidth of AGP. That means the PCI Express x16 slot is longer and has 16 PCI-Express interfaces to the slot, its kind of like if you ever saw the quad ethernet cards that were out when gigabit was really expensive. You could setup 4 100BaseT connections to act like one and run in parallel, effectively giving you a 400BaseT link with inexpensive technology.

PCI Express is a far supperior technology to PCI, yet it remains software compatible. Infact, PCI Express's serial and switching technologies are not even visible from the OS side of things, all PCI devices are configured the exact same way.

And of course, may PCI Express chipsets have PCI Express to PCI bridges, built in, allowing both PCI, PCI Express, and PCI Express x16 slots on the same machine.

As it is right now, PCI is overwealmed with data. Look at the XServe. 3 Serial ATA controllers running at 150MB/sec. 2 Gigabit ethernet controllers. An optional Fibre Channel card...lets not forget USB2 and FireWire 800. All of this is funnelled into the SAME PCI BUS. There is no way you can throttle all these devices at the same time. Now, what happens with PCI Express? You are now limited to the bottle necks in other parts of the system....... memory throughput to the chipset.... and CPU speed / front side bus speed....... both of which the PPC G5 platform are doing very well at.

I for one, am under strong beleif, that a new XServe will be announced in the August / September time frame with PCI Express, as it can take advantage of the PCI Express throughput the most with all the integrated devices in the server.

Wow...so in laymen's terms:

*Video card data that was originally on AGP (and thus part of the PCI bus), is now independent from the PCI bus? Thus giving more bandwidth to the remaining devices (USB/FW, ethernet, etc.)?

Or is it like a california highway, that switches from a 4-way stop, to a 4-lane (one direction) highway? Haha, I'm confusing myself now.

*And, b/c it is like a switch...no more waiting/latency bull that currently limits the PCI bus in nearly all computers today?

[fixyourthinking]

I was at WWDC working on the systems that had the new displays connected. I didn't see Tiger installed when I was using them.

Well, thanks - then it will be interesting see what Tiger can do for these machines with the optimized 64 bit code and new APIs, let alone optimized Nvidia drivers.

[shawnce]

just out of curiosity, how does one overclock a mac and how far can one go (without having to build in a prometeia mach 2 or something like that). You goto the store, buy a wall clock (or radio clock if you want to splurge), and put the clock on top of your Mac. Their is no step 4. :p

[ddtlm]

dguisinger:

As it is right now, PCI is overwealmed with data. Look at the XServe. 3 Serial ATA controllers running at 150MB/sec. 2 Gigabit ethernet controllers. An optional Fibre Channel card...lets not forget USB2 and FireWire 800. All of this is funnelled into the SAME PCI BUS.
No a lot of that is attached via other means, often integrated into chips connected to the HT bus. But an otherwise excellent post.

Its also noteworthy that a fair number of higher-end Machines (Suns, Opterons, big Intel machines) have multiple PCI busses.

[m.r.m.]

You goto the store, buy a wall clock (or radio clock if you want to splurge), and put the clock on top of your Mac. Their is no step 4. :p

not exactly a ruby this joke, but i'll give you credit for the effort! :p

[Roller]

I for one, am under strong beleif, that a new XServe will be announced in the August / September time frame with PCI Express, as it can take advantage of the PCI Express throughput the most with all the integrated devices in the server.

I was hoping, but not expecting, that the latest round of G5 desktops would have PCI Express. If XServe goes that route this summer or fall, though, I'd expect the PowerMac to do the same in the first quarter of 2005.

Although the real-world benefits of PCIE are limited today, that won't be so in a year from now.

[ktlx]

As it is right now, PCI is overwealmed with data. Look at the XServe. 3 Serial ATA controllers running at 150MB/sec. 2 Gigabit ethernet controllers. An optional Fibre Channel card...lets not forget USB2 and FireWire 800. All of this is funnelled into the SAME PCI BUS. There is no way you can throttle all these devices at the same time. Now, what happens with PCI Express? You are now limited to the bottle necks in other parts of the system....... memory throughput to the chipset.... and CPU speed / front side bus speed....... both of which the PPC G5 platform are doing very well at.

You know, for a post that started off telling someone else they did not know what they were talking about, it would help if you knew what you were talking about. None of the devices mentioned are on a shared PCI bus according to Apple's architecture diagram. Please see http://www.apple.com/xserve/architecture.html for more information.

[Mav451]

You know, for a post that started off telling someone else they did not know what they were talking about, it would help if you knew what you were talking about. None of the devices mentioned are on a shared PCI bus according to Apple's architecture diagram. Please see http://www.apple.com/xserve/architecture.html for more information.

You're joking me right? That picture is a diagram of the overall I/O connections in the X-serve--not specifically the PCI bus. This is why it shows the 2 G5's, memory, and FSB linked together (this is called the Northbridge) at the top.

At the bottom, is the Southbridge and its respective, connected devices. The PCI bus is part of the Southbridge--where the devices are connected to it: Hard drives (SATA or PATA, it doesn't matter); USB/FW800; and the Optical drives

I'm not sure how onboard Ethernet works, but it seems it has an "independent" bus for it on the XServe, so we might be able to factor it out of the equation (in terms of sharing the PCI bus bandwidth).

[MacinDoc]

just out of curiosity, how does one overclock a mac and how far can one go (without having to build in a prometeia mach 2 or something like that).
I don't know about the G5s, but my old beige G3 desktop has a series of jumpers that adjust the speed of the CPU and the system bus. When I put a G4 ZIF ugrade in my Mac, I had to change the jumpers for the new speed to be recognized (of course, this voided my long-expired Apple warranty). Later, after the warranty on my ZIF had expired, I played with the jumpers some more, increasing the speed of my 533 MHz ZIF to 583 MHz, and increasing the bus speed from 66 MHz to 83 MHz, all with no loss of stability (by the way, the original ZIF that the machine shipped with was a 233 MHz G3).

[dguisinger]

You know, for a post that started off telling someone else they did not know what they were talking about, it would help if you knew what you were talking about. None of the devices mentioned are on a shared PCI bus according to Apple's architecture diagram. Please see http://www.apple.com/xserve/architecture.html for more information.

Most of those devices are connected to the PCI bus, check the profiler.

Also, take a look at the original XServe (Which I own 2 of)...yes its a dual G4, so it doesn't have HT, but it had dual gig ethernet on PCI, along with 4 hard disk controllers + the ATA controller for the CD-ROM.

Yes the PCI bus cripples these machines.....of course PCI cripples intel based servers as well, so far Intel hasn't released their server chipset for PCI-E, but there is definately a need for the technology.

[dguisinger]

You're joking me right? That picture is a diagram of the overall I/O connections in the X-serve--not specifically the PCI bus. This is why it shows the 2 G5's, memory, and FSB linked together (this is called the Northbridge) at the top.

At the bottom, is the Southbridge and its respective, connected devices. The PCI bus is part of the Southbridge--where the devices are connected to it: Hard drives (SATA or PATA, it doesn't matter); USB/FW800; and the Optical drives

I'm not sure how onboard Ethernet works, but it seems it has an "independent" bus for it on the XServe, so we might be able to factor it out of the equation (in terms of sharing the PCI bus bandwidth).


Actually, since the other poster pointed out HT to the two ethernet controllers and told me to look it up because I'm stupid, I did:

Apple strengthens the Ethernet implementation on the Xserve G5 by adding a high-performance ASIC to the main logic board. This advanced microprocessor includes two independent 10/100/1000BASE-T Ethernet interfaces, each with its own interrupt, on a dedicated 64-bit, 133MHz PCI-X bus.

Looks like he's just as stupid. I didn't realize it was a seperate PCI bus, HOWEVER: Most multi-bus systems actually combine all the buses with PCI-PCI bridges, therefore grouping some bottlenecks, but ultimately access to the main CPU/memory still remains a bottle neck of the main PCI bus. Apple doesn't state how they designed their chipset.

[ddtlm]

dguisinger:

HOWEVER: Most multi-bus systems actually combine all the buses with PCI-PCI bridges, therefore grouping some bottlenecks, but ultimately access to the main CPU/memory still remains a bottle neck of the main PCI bus. Apple doesn't state how they designed their chipset.
Perhaps that was true once, but its not anymore. Apple's G5 PCI-X chip appears to be an AMD chip (AMD-8131) which sprouts no less that 3 PCI/PCI-X busses which are totally independent and connected, through the chip, to HT and from there into the system RAM through the northbridge (which is on-CPU for an AMD chip). The Sun machines I've speced recently have similarly indepent PCI busses.

http://www.amd.com/us-en/Processors/TechnicalResources/0,,30_182_739_9004,00.html

As far as ethernet, fireware, and so on... in PC-land I know these are very often integrated into chipsets and not connected to any PCI bus at all. I thought Apple was doing the same thing, seeing as how it is both more cost effective and faster.

[dguisinger]

dguisinger:


Perhaps that was true once, but its not anymore. Apple's G5 PCI-X chip appears to be an AMD chip (AMD-8131) which sprouts no less that 3 PCI/PCI-X busses which are totally independent and connected, through the chip, to HT and from there into the system RAM through the northbridge (which is on-CPU for an AMD chip). The Sun machines I've speced recently have similarly indepent PCI busses.

http://www.amd.com/us-en/Processors/TechnicalResources/0,,30_182_739_9004,00.html

As far as ethernet, fireware, and so on... in PC-land I know these are very often integrated into chipsets and not connected to any PCI bus at all. I thought Apple was doing the same thing, seeing as how it is both more cost effective and faster.

Umm, you do realize that the PCI bus exists within the Chipsets do you not? Most ethernet designs, like the ones on the X-serve are hooked into the internal PCI chipset, INTO the PCI bus. Want to know why? PCI is more than just sending data, its a method of configuration, a method of bus mastering, dma control, etc. HT doesn't provide that. HT provides a direct link from one point to another, its not designed to be a bus, its designed to be a point to point connection. The media kept talking about how HT was to be AMD's PCI replacement, going up against Intel's PCI Express (then known under a different name), and its funny, people beleived them. AMD was even quoted saying it was not designed to be a perhipheral expansion bus!

There are some systems that do support independent busses, HOWEVER: The PCI bus for example on x86 systems has a standard IO range. Putting a second bus on the same machine thats independent means any 3rd party OS will not detect it without being told where it is. Apple has a benifit in this situation because they control the platform, but not all architectures are as easily modifiable.

Have you read the PCI spec front to back? I haven't read every page of it, but the fact remains I have worked on kernel level PCI code. I have a book sitting on the shelf behind me that all it talks about is PCI... I have a fair deal of knowledge how PCI and PCI bridges / bus controllers work. Most devices will not ever connect directly into HT because it doesn't provide most of the functions devices need to actually communicate with the OS.

HT is like a wire.... PCI is the operator at the end of a TTD phone call.......it tells the PCI devices to configure how the OS wants them to..... As far as I can tell HT isn't configurable by the OS from what I've read, its just the line......a point to point serial protocol designed for higher front side bus speeds and to eliminate circuit trace complexity on the mother board. The OS still needs a standard interface for Plug-n-Play which PCI provides for detection, configuration, and operation.

[Mav451]

Yeah I remember...something along the lines of Hypertransport not competing with PCI Express, but working together (e.g. any of those Athlon 64 chipsets that will soon to be integrating PCI-E as a living testament to this).

A quick search found me this:
http://www.us.design-reuse.com/articles/article4199.html

Appropriately titled "How HyperTransport and PCI Express complement each other" :)

[ddtlm]

dguisinger:

There are some systems that do support independent busses, HOWEVER: The PCI bus for example on x86 systems has a standard IO range. Putting a second bus on the same machine thats independent means any 3rd party OS will not detect it without being told where it is.
What are you talking about? I'm talking about x86 machines with 3 PCI busses that I have handled with my own hands, no special configuration. They have full independent bandwidth. Lots of Opteron boards are like that. The Sun machines I speak of (V440 for example) also have full bandwidth, and the PCI busses aren't even hooked into the same chip on the system. I'm pretty sure some Itanium systems from HP have 6 or more PCI-X busses hooked to totally separate chips.

Have you read the PCI spec front to back? I haven't read every page of it, but the fact remains I have worked on kernel level PCI code. I have a book sitting on the shelf behind me that all it talks about is PCI... I have a fair deal of knowledge how PCI and PCI bridges / bus controllers work.
That's good, but your still making incorrect claims, such as:

HOWEVER: Most multi-bus systems actually combine all the buses with PCI-PCI bridges, therefore grouping some bottlenecks, but ultimately access to the main CPU/memory still remains a bottle neck of the main PCI bus.
Wrong, but don't ask me how to resolve the conflicts of the real world with the PCI spec cause I haven't a clue.

[melgross]

Guys, there's no point in arguing the fine points. It's fairly simple overall.

The PCI bus is finished. That's all that matters for Apple. PCI-X was invented as a cheaper lower technology competitor to the Intel invented Express bus. The only advantage to PCI-X is that as it is related to PCI 2, it is easy to engineer, and costs less. It has about twice the bandwidth and has been implemented by server companies who couldn't wait for Express to come out. Also, server companies are conservative, and wanted to go with something that wasn't re-engineered from scratch, which Express is.

Apple had no choice but to go with X because Express wasn't ready. It's just now coming out.

The industry estimates are that by January, 50% of all new Pc's will come with Express. However, a Pc being, after all, a Pc, a large number of those units will still have PCI 2.2 slots as well, ala PCI/ISA. Several mobo makers have also shown Express mobo's with 8x AGP on the board as well, though the performance is poor. It's just for compatibility.

Express ver. 1 is four times faster than PCI 2.2 about 1GB sec, vs. 266MB sec. Ver 2 arriving in 2006 will be about 2GB sec.

PCI-X is a dead end. Apple must shift to Express. The question is when. I'm hoping that January will be that time. Intel has shown their cards, and there were about a dozen at the show in, I forget now, I think Taiwan, last month.

If Apple doesn't do this soon, then they will suffer not being in step. There have been few X cards issued because most X makers have, as I've said, been Server companies who don't use most cards. What little there is will trickle to a halt sometime in 2005. Apple has to be at the front of this curve, not behind it. Whatever you may think of these busses, Apple has to be perceived as having the most muscle, and the greatest capabilities.

Having Express will help them to achieve that goal.

[faustofernos]

If you read this alound as the comic book guy (from the Simpsons) its hilarious! ha ha ha ha...

Umm, you do realize that the PCI bus exists within the Chipsets do you not? Most ethernet designs, like the ones on the X-serve are hooked into the internal PCI chipset, INTO the PCI bus. Want to know why? PCI is more than just sending data, its a method of configuration, a method of bus mastering, dma control, etc. HT doesn't provide that. HT provides a direct link from one point to another, its not designed to be a bus, its designed to be a point to point connection. The media kept talking about how HT was to be AMD's PCI replacement, going up against Intel's PCI Express (then known under a different name), and its funny, people beleived them. AMD was even quoted saying it was not designed to be a perhipheral expansion bus!

There are some systems that do support independent busses, HOWEVER: The PCI bus for example on x86 systems has a standard IO range. Putting a second bus on the same machine thats independent means any 3rd party OS will not detect it without being told where it is. Apple has a benifit in this situation because they control the platform, but not all architectures are as easily modifiable.

Have you read the PCI spec front to back? I haven't read every page of it, but the fact remains I have worked on kernel level PCI code. I have a book sitting on the shelf behind me that all it talks about is PCI... I have a fair deal of knowledge how PCI and PCI bridges / bus controllers work. Most devices will not ever connect directly into HT because it doesn't provide most of the functions devices need to actually communicate with the OS.

HT is like a wire.... PCI is the operator at the end of a TTD phone call.......it tells the PCI devices to configure how the OS wants them to..... As far as I can tell HT isn't configurable by the OS from what I've read, its just the line......a point to point serial protocol designed for higher front side bus speeds and to eliminate circuit trace complexity on the mother board. The OS still needs a standard interface for Plug-n-Play which PCI provides for detection, configuration, and operation.

[ktlx]

You're joking me right? That picture is a diagram of the overall I/O connections in the X-serve--not specifically the PCI bus. This is why it shows the 2 G5's, memory, and FSB linked together (this is called the Northbridge) at the top.

At the bottom, is the Southbridge and its respective, connected devices. The PCI bus is part of the Southbridge--where the devices are connected to it: Hard drives (SATA or PATA, it doesn't matter); USB/FW800; and the Optical drives

I'm not sure how onboard Ethernet works, but it seems it has an "independent" bus for it on the XServe, so we might be able to factor it out of the equation (in terms of sharing the PCI bus bandwidth).

Did you even bother to read the page I referenced? Either you did not or you did not understand what it is saying.

[Mav451]

You know, for a post that started off telling someone else they did not know what they were talking about, it would help if you knew what you were talking about. None of the devices mentioned are on a shared PCI bus according to Apple's architecture diagram. Please see http://www.apple.com/xserve/architecture.html for more information.

Actually I did. This time, I'll even circle the picture for you.
(top red square = northbridge, lower red square = southbridge)
(blue circles = devices connected to southbridge, via PCI bus and talk to the devices in the Northbridge by HyperTrasnport--think of it as a wire between the Northbridge and Southbridge)

Upon further inspection, at least by the diagram, it seems the Ethernet and PCI-X are independent of the PCI bus (as they are not connected to the HyperTransport at the bottom).

To answer your question, ALL of the devices he mentioned, except the Gigabit Ethernet, are funneled through the PCI bus:


As it is right now, PCI is overwealmed with data. Look at the XServe. 3 Serial ATA controllers running at 150MB/sec. 2 Gigabit ethernet controllers. An optional Fibre Channel card...lets not forget USB2 and FireWire 800.


As I indicated on my picture, the SATA and USB2/FW800 are both linked to the HT (and consequently are part of the PCI bus, run BY the southbridge). The Fiber Channel cards, it seems, are normal PCI cards, NOT PCI-X cards, so they do not get the advantage of running independent of the PCI bus (and hence, are part of the latency or "wait" situation).

[shawnce]

Guys why are you working from an artitist rendering and market speak? Go to the real documentation (well as real as Apple publicly releases).

http://developer.apple.com/documentation/Hardware/Developer_Notes/Macintosh_CPUs-G5/PowerMacG5/ (http://developer.apple.com/documentation/Hardware/Developer_Notes/Macintosh_CPUs-G5/PowerMacG5)

In particular the chapters on block diagram (http://developer.apple.com/documentation/Hardware/Developer_Notes/Macintosh_CPUs-G5/PowerMacG5/2Architecture/chapter_3_section_2.html#//apple_ref/doc/uid/TP40001397-CH206-TPXREF102) and K2 I/O Controller (http://developer.apple.com/documentation/Hardware/Developer_Notes/Macintosh_CPUs-G5/PowerMacG5/2Architecture/chapter_3_section_7.html#//apple_ref/doc/uid/TP40001397-CH206-TPXREF115).

Note the following diagram contained in that documentation...

http://developer.apple.com/documentation/Hardware/Developer_Notes/Macintosh_CPUs-G5/PowerMacG5/art/040177001721_01c.gif

Also such data exists for the Xserve as well (rather close to the current G5s)...

http://developer.apple.com/documentation/Hardware/Developer_Notes/Servers/XserveG5/index.html

http://developer.apple.com/documentation/Hardware/Developer_Notes/Servers/XserveG5/art/031224001708_01.gif

[Mav451]

Good god...that's quite a find.

[shawnce]

FYI by my math all of the IO devices connected to the K2 I/O controller (south bridge) cannot peak much above 800 MBps (megabytes per second). The HT bus feeding the K2 can support up to 1600 MBps (full duplex).

Also note that the K2 internally appears to simulate many PCI buses (not one big shared PCI bus) with PCI bridges off of those to interface to the i2s, SATA, etc. buses supported (as seen with IOReg). It does expose one PCI 64b/33MHz bus for attaching the USB controller, airport card, and boot ROM.

Anyway I do not believe that all devices attached to the K2 I/O controller share a single PCI bus internally but the K2 supports many internal PCI buses (possibly simulated for the benefit of software drivers) all bridged off of the one HT bus attaching it to the system. An HT bus with enough bandwidth to cover any peak bandwidth loads. The chip set is well balanced with little to no bottlenecks designed into it.

[stuepfnick]

The 1.6 GBs throuput seems to suffice for all that stuff.

450 MBs for 3 HDs
050 MBs for an extreme optical drive
100 MBs FW800
100 MBs 2x FW400
150 MBs 3x USB 2.0
001 MBs bluetooth
005 MBs Airport Extreme
120 MBs Gigabit Ethernet
010 MBs Audio stuff

The PCI slots are seperated, so that makes totalle 986 MBs throughput! The maximum is 800 MBs in one direction. And you will never ever in any way get full usage of all this stuff at the same time! That would mean max. possible read speed of all drives, of all I/O devices, etc. This is really impossible in any way of usage you can think of.
So can't see any bottleneck there in any way...

[shawnce]

Good god...that's quite a find.
Apple has always provided documentation such as this on their developer site (no registration needed). :)

http://developer.apple.com/documentation/Hardware/AppleHardware-date.html

Note the "All Hardware Products" at the bottom to see some real history and just how far they have come.

[stuepfnick]

The 1.6 GBs throuput seems to suffice for all that stuff.

450 MBs for 3 HDs
050 MBs for an extreme optical drive
100 MBs FW800
100 MBs 2x FW400
150 MBs 3x USB 2.0
001 MBs bluetooth
005 MBs Airport Extreme
120 MBs Gigabit Ethernet
010 MBs Audio stuff

The PCI slots are seperated, so that makes totalle 986 MBs throughput! The maximum is 800 MBs in one direction. And you will never ever in any way get full usage of all this stuff at the same time! That would mean max. possible read speed of all drives, of all I/O devices, etc. This is really impossible in any way of usage you can think of.
So can't see any bottleneck there in any way...

Oh sorry, the Ethernet is also seperated, so that could actually only get to the limit, when all I/O connectors run at full speed in the same direction, which is 100% impossible.

[gossas]

There's obviously naff all going on as people are now arguing over diagramatical and mathematical theoretical values. Bottom line, until someone can guide me through using Macromedia suite on my new machine and I subjectively believe that everything runs faster, then why should I believe it's faster. On paper all new machines are faster or slower by design, benchmarks quantify theory and real world tests are subjective and qualitative. No test has any bearing on my enjoyment of my new machine.

[fregedegpo]

My big wish is that the sound issues will be solved that came from the power supply corrupting the anolog audio outs. I bought a dual2 G5 for myself and the problem was aweful - luckily the store agreed to return the unit - and then I bought one for my work and found the same problem. 2 out of 2 is not good!

I need a new cpu so badly, as I use Logic on my dual500 G4 and the new plugins can barely make it on these procs. But I don't want to buy a computer that has serious sound problems. Does anybody know about the dual 2.5 G5, or even these revised dual 2 G5s?

Thank you folks, I hope somebody has some info.... :)

[nuckinfutz]

My big wish is that the sound issues will be solved that came from the power supply corrupting the anolog audio outs. I bought a dual2 G5 for myself and the problem was aweful - luckily the store agreed to return the unit - and then I bought one for my work and found the same problem. 2 out of 2 is not good!

I need a new cpu so badly, as I use Logic on my dual500 G4 and the new plugins can barely make it on these procs. But I don't want to buy a computer that has serious sound problems. Does anybody know about the dual 2.5 G5, or even these revised dual 2 G5s?

Thank you folks, I hope somebody has some info.... :)

You're "never" going to want to use the audio capabilities of the computer(hardware wise). After getting Logic you'll need to investigate a good audio interface. PCI or FW or USB, Core Audio was meant to handle these quite well.

[gerardrj]

Here's another table of G5 optimized CineBench numbers for the 2.0 vs 2.5 hz models:

1st gen Dual G5 @ 2GHz; 1.5GB RAM; stock Radeon 9600w/64MB
Supplied 2x2.5 numbers in parenthesis ()
calculated percent change in brackets []

Rendering (Single CPU): 287 (356) [+24%] CB-CPU
Rendering (Multiple CPU): 510 (633) [+24%] CB-CPU
Multiprocessor Speedup: 1.78 (1.78) [0%]

Shading (CINEMA 4D) : 278 (335) [+20%] CB-GFX
Shading (OpenGL Software Lighting) : 776 (995) [+28%] CB-GFX
Shading (OpenGL Hardware Lighting) : 1507 (1794) [+19%] CB-GFX
OpenGL Speedup: 5.43 (5.36) [-2%]

Apparently the 2.5Ghz machine achieves little to no performance gain because of the smaller process size, the overall difference in performance is about 25%, sometimes a little higher, sometimes a little lower but it equates solely to clock speed increase.

[eric_n_dfw]

Apparently the 2.5Ghz machine achieves little to no performance gain because of the smaller process size, the overall difference in performance is about 25%, sometimes a little higher, sometimes a little lower but it equates solely to clock speed increase.Forgive my ignorance, but why would the smaller process size in and of itself increase performace? I thought the point of the smalller size was to shorten the distances the electrons need to travel, thus allowing the higher clock speeds.

[veggyrites]

didn't the 30" use two GPU's to run, on the keynote Steve showed us half a screen with only one GPU.

that is all :confused:

[DWKlink]

didn't the 30" use two GPU's to run, on the keynote Steve showed us half a screen with only one GPU.

that is all :confused:

I believe that its only one card. The half screen was a limitation of the DVI interface. A single DVI connection cannot carry all the information needed to drive the monitor at full rez. So they developed Dual DVI - basically two DVI paths on one cable running in parallel as I understand it - to drive the monitor. And the 6800 includes 2 of these dual DVI ports to drive up to 2 30" displays at full rez. But I think that the card itself only includes one GPU.

[veggyrites]

ok , that is better

[eric67]

Here's another table of G5 optimized CineBench numbers for the 2.0 vs 2.5 hz models:

1st gen Dual G5 @ 2GHz; 1.5GB RAM; stock Radeon 9600w/64MB
Supplied 2x2.5 numbers in parenthesis ()
calculated percent change in brackets []

Rendering (Single CPU): 287 (356) [+24%] CB-CPU
Rendering (Multiple CPU): 510 (633) [+24%] CB-CPU
Multiprocessor Speedup: 1.78 (1.78) [0%]

Shading (CINEMA 4D) : 278 (335) [+20%] CB-GFX
Shading (OpenGL Software Lighting) : 776 (995) [+28%] CB-GFX
Shading (OpenGL Hardware Lighting) : 1507 (1794) [+19%] CB-GFX
OpenGL Speedup: 5.43 (5.36) [-2%]

Apparently the 2.5Ghz machine achieves little to no performance gain because of the smaller process size, the overall difference in performance is about 25%, sometimes a little higher, sometimes a little lower but it equates solely to clock speed increase.

good info.
hardmac.com is also providing very similar resutls, one should have a look.
http://www.hardmac.com/niouzcontenu.php?date=2004-07-05#2434

[WinX]

It is obvious you don't understand what PCI Express is. PCI Express is (AFAIK) a 4-wire RX/TX pair capable of pushing in serial more bandwidth than a normal PCI slot, and if I remember correctly it is a jump from 133MB/sec to 150MB. Someone can correct me on that if I'm wrong.Correct you? Easy. A 1x PCI Express is clocked at 2.5 Gigabits/sec...


On standard PCI, ALL SLOTS SHARE the 133MB/sec bus!On an antiquated decade-old PC, maybe. What you don't seem to understand is that a modern Southbridge chip can implement a multitude of INDEPENDENT PCI buses.


As it is right now, PCI is overwealmed [sic] with data. Look at the XServe. 3 Serial ATA controllers running at 150MB/sec. 2 Gigabit ethernet controllers. An optional Fibre Channel card...lets not forget USB2 and FireWire 800. All of this is funnelled into the SAME PCI BUS.Um, the operative word, even on a G4 Xserve, is INDEPENDENT buses. Serial ATA disks, Firewire and USB interfaces DO NOT need to communicate with each other, and thus DO NOT need to SHARE a bus.
The computer-side end point of an I/O operation will typically be RAM -- e.g. via a DMA controller -- or a CPU. The I/O links are thus conceptually point-to-point; Apple's G5-based machines implement that point-to-point paradigm using a combination of hardware -- e.g. the G5's Apple-designed and IBM-manufactured "Northbridge" -- and software -- multiplexing of logical point-to-point links over a single 800 Megabytes/sec multiplxed Hypertransport channel to the K2 controller.

We should thank stuepfnick for showing in a previous post that it would be next to impossible to overwhelm an Xserve G5's 800 MB/s HT channel in real-world conditions...


I for one, am under strong beleif, that a new XServe will be announced in the August / September time frame with PCI Express, as it can take advantage of the PCI Express throughput the most with all the integrated devices in the server.Um, the 2 Gigabytes/sec throughput offered by the G5 Xserve's PCI-X slots is sufficient to comfortably support even cards requiring high throughputs. Fiberchannel interfaces run at 1.06 to 2.12 Gigabits/sec. Infiniband runs at 10 Gigabits/sec with 8b/10b encoding, thus actual peak payload throughput is more like 8 Gigabits/sec.


Have you read the PCI spec front to back? I haven't read every page of it, but the fact remains I have worked on kernel level PCI code.Um, PCI semantics do not necessarily need a physical parallel "PCI" bus. The PCI protocol can be implemented as a layer above a serial physical layer, like PCI-Express or Hypertransport. People unfamiliar with modern hardware design concepts or point-to-point traffic patterns apparently don't realize that PCI driver software wouldn't be able to tell the difference.

[g4cubed]

I believe that its only one card. The half screen was a limitation of the DVI interface. A single DVI connection cannot carry all the information needed to drive the monitor at full rez. So they developed Dual DVI - basically two DVI paths on one cable running in parallel as I understand it - to drive the monitor. And the 6800 includes 2 of these dual DVI ports to drive up to 2 30" displays at full rez. But I think that the card itself only includes one GPU.


dual link DVI passes along twice the information of standard DVI along the same cable by using all 24 pins of the connector instead of just 12. All this graphics power comes at a space price as well—the card itself is so big that it takes up the space of two PCI cards

[Soire]

Any idea why my 2.5 order would have a shipping date of 8/2 if i ordered it only a couple of days after release?

It is a BTO, but all I added was a bigger hard drive, AE, and Bluetooth. Oh, and it was ordered as a developer. Any ideas when it may REALLY ship?

[MacinDoc]

Forgive my ignorance, but why would the smaller process size in and of itself increase performace? I thought the point of the smalller size was to shorten the distances the electrons need to travel, thus allowing the higher clock speeds.
Smaller process size allows for higher theoretical clock speeds, but does not increase the processor speed at any given clock speed. A 2 GHz processor completes 2 billion cycles per second, whether it is on a 130 nm, a 90 nm, or a 65 nm process.

[gerardrj]

Forgive my ignorance, but why would the smaller process size in and of itself increase performace? I thought the point of the smalller size was to shorten the distances the electrons need to travel, thus allowing the higher clock speeds.

Shorter distance for electrons to travel = less latency = higher performance.
It's not a lot of performance, but it could be measurable under the right circumstances.

[pounce]

Any idea why my 2.5 order would have a shipping date of 8/2 if i ordered it only a couple of days after release?

It is a BTO, but all I added was a bigger hard drive, AE, and Bluetooth. Oh, and it was ordered as a developer. Any ideas when it may REALLY ship?

my BTO 2.5 was ordered from the applestore online and still shows on or by 7/30 as the shipping. it was hinted that it may arrive sooner and that the given date was a safe estimate. i'd love to know if some of these would really be shipping sooner. any ideas?

[zac4mac]

I got bummed when I saw the numbers. My poor old slow Rev A G5. Pffffft.
Ran C4D-G5Beta myself, 1.5GB RAM

CINEBENCH 2003 v1
****************************************************

Tester : Zack

Processor : DP G5
MHz : 2 x 2.0 GHz
Number of CPUs : 2
Operating System : 10.2.4

Graphics Card : 9800Pro OEM 128MB 22"ACD

****************************************************

Rendering (Single CPU): 288 CB-CPU 356
Rendering (Multiple CPU): 510 CB-CPU 633

Multiprocessor Speedup: 1.77 1.78

Shading (CINEMA 4D) : 276 CB-GFX 335
Shading (OpenGL Software) : 797 CB-GFX 995
Shading (OpenGL Hardware) : 1526 CB-GFX 1794

OpenGL Speedup: 5.52 5.36


Looks like a decent speed bump, but not a terribly huge one.

Z

[fregedegpo]

You're "never" going to want to use the audio capabilities of the computer(hardware wise). After getting Logic you'll need to investigate a good audio interface. PCI or FW or USB, Core Audio was meant to handle these quite well.
Yes, of course! I already use the motu828 and will hopefully soon be purchasing a Metric Halo. I still found that the computer was still feeding my mixer noise. Hmm... what was I doing wrong?

[DWKlink]

Any idea why my 2.5 order would have a shipping date of 8/2 if i ordered it only a couple of days after release?

It is a BTO, but all I added was a bigger hard drive, AE, and Bluetooth. Oh, and it was ordered as a developer. Any ideas when it may REALLY ship?

I was under the impression that any BTO machine would take longer to deliver, because it has to be customized. eg - a current 1.8 DP that is ordered off the apple store has a ship date of the same business day. If you add a bigger drive, blue tooth and AE, the ship date moves up to 7-10 business days. And considering that they have to get the machine in stock (presumably the end of july) before they can customize it, that might explain the delay in your ship date until early august.

If you keep your fingers crossed like me though, maybe apple is just being conservative in their expected ship dates and we'll all get them early.

[isgoed]

I got bummed when I saw the numbers. My poor old slow Rev A G5. Pffffft.
Ran C4D-G5Beta myself, 1.5GB RAM

CINEBENCH 2003 v1
****************************************************

Tester : Zack

Processor : DP G5
MHz : 2 x 2.0 GHz
Number of CPUs : 2
Operating System : 10.2.4

Graphics Card : 9800Pro OEM 128MB 22"ACD

****************************************************

Rendering (Single CPU): 288 CB-CPU 356
Rendering (Multiple CPU): 510 CB-CPU 633

Multiprocessor Speedup: 1.77 1.78

Shading (CINEMA 4D) : 276 CB-GFX 335
Shading (OpenGL Software) : 797 CB-GFX 995
Shading (OpenGL Hardware) : 1526 CB-GFX 1794

OpenGL Speedup: 5.52 5.36


Looks like a decent speed bump, but not a terribly huge one.

Z

Yeah you got the radeon 9800pro and 1.5Gb Ram.

This is at least a fair comparisson between the old high-end and the new high-end. Processor speed is up by 25% as expected and OpenGL with 20%, which seems a bit low, so either nvidia or cinebench should put more effort in harnassing the power of the gefore 6800. It would be nice to see also a bench with the 2.5Ghz and the radeon 9800 to see wether it is worth the money to buy the gefore 6800.

[pounce]

Yes, of course! I already use the motu828 and will hopefully soon be purchasing a Metric Halo. I still found that the computer was still feeding my mixer noise. Hmm... what was I doing wrong?

this is being discussed now on either motu mac or daw mac email lists. one thing mentioned is actually disconnecting the cable in the computer that relates to it's internal sound. that seemed to be where the noise came in from. some sort of grounding issue is suspected. after that some users reported their interfaces were dead quiet again. see if you can do that and let us know how it works for you.

[fregedegpo]

this is being discussed now on either motu mac or daw mac email lists. one thing mentioned is actually disconnecting the cable in the computer that relates to it's internal sound. that seemed to be where the noise came in from. some sort of grounding issue is suspected. after that some users reported their interfaces were dead quiet again. see if you can do that and let us know how it works for you.
Will do, buddy! But I have to buy the computer first! Again, I was hoping to find out if the new cpus are still having this problem in the first place. Nothing would make me happier than going all digital. I just need a new mixing board for that... :)

[JackAxe]

How does one use two 30" screens on a G5, if just one requires both ports on the 6800? What am I missing? :)

<]=)

[pounce]

How does one use two 30" screens on a G5, if just one requires both ports on the 6800? What am I missing? :)

<]=)

each 30" screen uses one dvi port. here is the link...

http://www.apple.com/displays/digital.html

read the dvi section

[JackAxe]

That's the part that was confusing me. I thought that "Dual Link DVI" meant that 2 cables were required like the Viewsonic 2290b. I just found the info in their tech-spec PDF though that talks about how one DVI port can use a single or dual-link bandwidth. So I'm no longer confused. :)

<]=)

[eric_n_dfw]

These two answers contradict each other:

Smaller process size allows for higher theoretical clock speeds, but does not increase the processor speed at any given clock speed. A 2 GHz processor completes 2 billion cycles per second, whether it is on a 130 nm, a 90 nm, or a 65 nm process.

Shorter distance for electrons to travel = less latency = higher performance.
It's not a lot of performance, but it could be measurable under the right circumstances.

Am I wrong?

[g4cubed]

These two answers contradict each other:

Quote:
Originally Posted by MacinDoc
Smaller process size allows for higher theoretical clock speeds, but does not increase the processor speed at any given clock speed. A 2 GHz processor completes 2 billion cycles per second, whether it is on a 130 nm, a 90 nm, or a 65 nm process.

Quote:
Originally Posted by gerardrj
Shorter distance for electrons to travel = less latency = higher performance.
It's not a lot of performance, but it could be measurable under the right circumstances.



Am I wrong?

No, they don't.
To my knowledge they're both saying similar things. The smaller size (90nm) does mean a shorter distance for electrons to travel = less latency (response time for the electrons to move) = higher performance.
Macindoc states this as "Smaller process size allows for higher theoretical clock speeds" which is also true.
Just making it smaller doesn't make it that much faster. Other things are involved for you to obtain higher clock speeds. Smaller also means lower power usage.

[gekko513]

No, they don't.
To my knowledge they're both saying similar things. The smaller size (90nm) does mean a shorter distance for electrons to travel = less latency (response time for the electrons to move) = higher performance.
Macindoc states this as "Smaller process size allows for higher theoretical clock speeds" which is also true.
Just making it smaller doesn't make it that much faster. Other things are involved for you to obtain higher clock speeds. Smaller also means lower power usage.
Actually making it smaller doesn't make it faster at all even if the latency, as described here, is smaller.

A smaller latency would theoretically allow the cpu to be run at higher clock, if other issues still prevents the cpu from being stable at higher clock speeds, then you gain nothing from the lower latency.

You could view this as measurable voltage traveling between transistors. In a smaller cpu, the voltage needs to travel a shorter distance (less latency if you like), but the voltage is only measured at every clock tick, so even if the voltage state arrived earlier at the next transistor gate with a die shrink, the voltage will not be measured earlier if the clock frequency isn't increased.

[aussiemac86]

What are peoples due dates for non-BTO dual 2.5's?
especially anyone in Aus. Mine is ordered through a reseller so i was wondering how my waiting time compares to others.

[JackAxe]

That's a massive jump aussiemac86. You should get more RAM though. :)

<]=)

[MacinDoc]

No, they don't.
To my knowledge they're both saying similar things. The smaller size (90nm) does mean a shorter distance for electrons to travel = less latency (response time for the electrons to move) = higher performance.
Macindoc states this as "Smaller process size allows for higher theoretical clock speeds" which is also true.
Just making it smaller doesn't make it that much faster. Other things are involved for you to obtain higher clock speeds. Smaller also means lower power usage.
I believe you are correct. It is the shorter distance travelled by electrons that allows the processor to complete more cycles per second. If you're running around a 90m track, you will complete more laps in ten minutes than if you were running around a 130m track. I doubt that a 130 nm chip could remain stable running at 2.5 GHz, just as you would be more fatigued trying to complete the same number of laps of the longer track in the same amount of time.

The clock speed is an objective measure of how many cycles a processor completes in 1 second. This is not, however, the same as the number of calculations per second, because modern processors complete more than one calculation per second, and the difference in the number of calculations per cycle is one of the reasons that a PPC 970 running at 2.5 GHz does not have the same apparent speed as a Pentium 4 chip running at 2.5 GHz. From what I understand, however, the architecture (number of VPUs etc) of the 970 is the same as the 970FX, so they should complete the same number of calculations per cycle.

Of course, take all of this with a grain of salt, because I am neither a physicist nor an engineer... ;)

[Soire]

That's a massive jump aussiemac86. You should get more RAM though. :)

<]=)

I'll be going from a g3 500 Mhz iBook with I believe 67 Mhz Frontside bus to a 2.5 G5 DualProcessor with 9600XT and 1.25 Ghz frontside bus. Not as big a jump as aussiemac, but still 5x is gonna blow me away... :)

[MacinDoc]

I'll be going from a g3 500 Mhz iBook with I believe 67 Mhz Frontside bus to a 2.5 G5 DualProcessor with 9600XT and 1.25 Ghz frontside bus. Not as big a jump as aussiemac, but still 5x is gonna blow me away... :)
Stop! You're both making me jealous! ;)

[barefeats]

The G5/2.0GHz MP Cinebench rating listed in comparison with the G5/2.5GHz MP doesn't appear to be "correct." I've used the "G5 beta" version of Cinebench 2003 on a G5/2.0GHz MP with Radeon 9800 OEM graphics.

I get a CPU rating of 522, making the G5/2.5 only 21% faster.

I get an OpenGL HardWare rating of 1505.