Xeon vs i7?

[Tesselator]

Y....


If i was rich (which im not)

If... So that's a "no" then.

 

[Ploki]

but if you consider what MS actually has to do to get it to work on so much different hardware, it's amazing it actually works.

i just love this statement its so true. and kind of not true

brr. if i remember the times when i did a clean reinstall of a windows box. (95,98,xp)
graphics at 800x600 no sound and even keyboard was barely working he he.
so while it does work, it works just barely. i dont know how is it with vista
3 cheers for 3rd party driver manufacturers.

 

[davewolfs]

Aren't the new Nehalem based Xeon's virtually the same as the i7's? Aside from the CPU cache and use of ECC memory, how are these processors any different?

 

[Tallest Skil]

Aren't the new Nehalem based Xeon's virtually the same as the i7's? Aside from the CPU cache and use of ECC memory, how are these processors any different?

The 3500 series is a rebranded Bloomfield with ECC turned on.

Gainestown is markedly different.

 

[Outsider]

I heard a nasty rumor that even the 5500 is essentially the same as the 3500 and i7. That the 3500 and i7 also have a second QPI bus but it is disabled (permanently, so no chance of re-enabling).

 

[davewolfs]

The 3500 series is a rebranded Bloomfield with ECC turned on.

Gainestown is markedly different.

Aside from having the ability to work in tandem with another processor, what is markedly different between the two aside from the high price tag?

If you don't have an application that can't max out more then 4 CPU's I'd be hard pressed to believe that the 2.26 Octo is any faster then the 2.66 Quad.

 

[MH01]

Dude, I built my own pcs for years, Im a former gamer. I played in Cal. -i (Cyber-athlete amateur league invitational ) I researched my mac for months before I bought it, I understood that I was making a sacrifice regarding GPU power but I expected the high powered CPU to compensate for it. Also these guys usually don't research anything, trust me If you ask any of them about there computer on vent they will tell you they bought whatever was the latest and greatest for the money. Also I am yet to find an 8-core pc gamer.

First of all there is no benefit in 8-core gaming (waste of money), 2 core is plenty. Being a gamer you should know that CPU cannot compensate for a weak GPU. To get a very good gaming machine you must eliminate bottlenecks. That being said, it all depends on what games you play. If you were playing counterstrike etc... big deal anything can run that.

I should probably take a step back and say that i am referring to running the newest releases at highest levels, that is why i refer to macs as lacking. If you play older less graphic intensive games, then a mac may be absolutely fine. My point is that no top range mac will ever beat a top range PC in gaming.

Just to fill you in, I have an imac 3.06 but I would never get rid of my gaming PC.

 

[MH01]

I think that you are confusing gamers with your average PC user, most of whom don't game.
While I do think that most gamers actually choose to use PC's I think the majority use PC's because that is what is used in their workplaces, and most if not everybody they know use PC's.
Also as for your argument that mac users are less IT literate than PC users, in general this is just not true. If anyone in my family has a problem with their PC I'm the one the ask to fix it.
Plus mac users are the ones to know enough about IT to know that there is another way.

I agree with you, I do not think any of us have stats/evidence to show what an average user is. Given most users are PC users, yeah by default that would bring the average down.

I am confident that Mac users get more out of their systems then PC users. OS X just works better for peoples simple computer needs.

 

[apolloa]

That's pathetic. He only needs to pull up a link to an Intel development release to get you to stop being childish.

From now on, let's all preempt that by not being childish.

Well, WHY has no one posted a link yet then genius? And it's not being childish what so ever, if YOUR stating as FACT something, back it up, I don't see any proof from you on anything you say. And how old are you? 35? Worked in IT for 8 years? Long enough to know that Xeons have alway's had extra arcitechture on them for additional instructions and processing? Hmm, though not.
I want a document from Intels website detailing exactly how the 5500 is no different to the i7.

 

[Tesselator]

I heard a nasty rumor that even the 5500 is essentially the same as the 3500 and i7. That the 3500 and i7 also have a second QPI bus but it is disabled (permanently, so no chance of re-enabling).

If they are that's a good thing! The technology screeeeems!

The only bummer is that anyone using the 3500 series in their designs or purchasing one is getting ripped off BIG TIME! As the >$2000 3500 series systems will have no computing advantages over the much more inexpensive Corei7 systems.

 

[Outsider]

If they are that's a good thing! The technology screeeeems!

The only bummer is that anyone using the 3500 series in their designs or purchasing one is getting ripped off BIG TIME! As the >$2000 3500 series systems will have no computing advantages over the much more inexpensive Corei7 systems.

Well that's true in a sense if you discount the memory support features. Even the prices between the 3500 and i7 are the same.

 

[sneezymarble]

Well, WHY has no one posted a link yet then genius? And it's not being childish what so ever, if YOUR stating as FACT something, back it up, I don't see any proof from you on anything you say. And how old are you? 35? Worked in IT for 8 years? Long enough to know that Xeons have alway's had extra arcitechture on them for additional instructions and processing? Hmm, though not.
I want a document from Intels website detailing exactly how the 5500 is no different to the i7.

Well, whatever fairy dust magic you think Xeons have over and above their desktop counterparts, the magic hasn't, at least in the case of C2D vs Xeons, proven to grant any performance advantage. I ran some encode, render, photoshop, and Geekbench benchmarks a while back on a single processor quad core Mac Pro 2.8GHz. My cheap little Q6600, when clocked at 2.8GHz, performed pretty much identically.

 

[Outsider]

It seems that the 5500 chipset (not Xeon X5500 series) is fairly expensive and it may be shared across the entire Mac pro line, leading to the cost increase of the low end Mac Pro. It seems as if the 5500 chipset can support the Xeon 3500 series provided you use only one processor.

 

[Tesselator]

There are of course benefits to 8-core computing. Here's just one example:

Lightwave3D 8-core render 1024x720 at 1000 frames: 52sec. per frame, Total = 14 hours.
Lightwave3D 2-core render 1024x720 at 1000 frames: 270sec. per frame, Total = 3.125 Days (75 hours).

Photoshop 3 levels of Lens Blur per frame same 1000 frames:
8-Core 2.66 = 32sec. per frame, Total = 8.9 hours
2-Core 2.66 = 108sec. per frame, Total = 1.25 Days (30 hours)

Upsample, downsample and encode all three image sequences to 2 different video (codec) formats:
8-Core: 3.33 Hours Total
2-Core: 10.4 Hours Total

That's just one project and yes it was actually timed. I saved about 87 hours of computing time - my paid time! That's 3.6 days and at $500 a day I would have to say that's a pretty massive benefit!!! Especially considering that this advantage is realized every time I do a job. If I do 5 of those jobs a month my 8 core system is making me an extra $10,000 every month!

So, no one can say there's no benefit. Even some games now have multi-core support. Quake 4 for example!

 

[Nik]

Now, what was this nonsense about no benefit?

Excuse me, he ment for gaming purposes.

 

[robinp]

LOL!!!

Lightwave3D 8-core render 1024x720 at 1000 frames: 52sec. per frame, Total = 14 hours.
Lightwave3D 2-core render 1024x720 at 1000 frames: 270sec. per frame, Total = 3.125 Days (75 hours).

Photoshop 3 levels of Lens Blur per frame same 1000 frames:
8-Core 2.66 = 32sec. per frame, Total = 8.9 hours
2-Core 2.66 = 108sec. per frame, Total = 1.25 Days (30 hours)

Upsample, downsample and encode all three image sequences to 2 different video (codec) formats:
8-Core: 3.33 Hours Total
2-Core: 10.4 Hours Total

That's just one project and yes it was actually timed. I saved about 87 hours of computing time - my paid time! That's 3.6 days and at $500 a day I would have to say that's a pretty massive benefit!!! Especially considering that this advantage is realized every time I do a job. If I do 5 of those jobs a month my 8 core system is making me an extra $10,000 every month!

Now, what was this nonsense about no benefit?

I'm absolutely not disagreeing with you, but the person you were replying to was talking about games, not 3D content creation....

edit - dude, you've just changed the quote to skip the bit about gaming... really rather uncool

 

[Tesselator]

Well that's true in a sense if you discount the memory support features. Even the prices between the 3500 and i7 are the same.

What do you mean? What are the memory support features you're speaking of?

 

[Tesselator]

I'm absolutely not disagreeing with you, but the person you were replying to was talking about games, not 3D content creation....

Yeah, I know. I just wanted to clarify what 8-cores were good for and I happen to have those numbers in front of me, so...

EDIT: dude, you've just changed the text to add the bit about the changed quote... really rather uncool (LOL! )

 

[robinp]

Yeah, I know. I just wanted to clarify what 8-cores were good for and I happen to have those numbers in front of me, so...

your numbers are interesting, although you are largely preaching to the converted. I think everyone realises that for 3D rendering number of cores scales performance almost linearly (90-95%)

 

[robinp]

Yeah, I know. I just wanted to clarify what 8-cores were good for and I happen to have those numbers in front of me, so...

EDIT: dude, you've just changed the text to add the bit about the changed quote... really rather uncool (LOL! )

haha yes, and as you noticed I clearly put my changes after 'edit' rather than changing a quote.....

 

[MH01]

LOL!!!

Lightwave3D 8-core render 1024x720 at 1000 frames: 52sec. per frame, Total = 14 hours.
Lightwave3D 2-core render 1024x720 at 1000 frames: 270sec. per frame, Total = 3.125 Days (75 hours).

Photoshop 3 levels of Lens Blur per frame same 1000 frames:
8-Core 2.66 = 32sec. per frame, Total = 8.9 hours
2-Core 2.66 = 108sec. per frame, Total = 1.25 Days (30 hours)

Upsample, downsample and encode all three image sequences to 2 different video (codec) formats:
8-Core: 3.33 Hours Total
2-Core: 10.4 Hours Total

That's just one project and yes it was actually timed. I saved about 87 hours of computing time - my paid time! That's 3.6 days and at $500 a day I would have to say that's a pretty massive benefit!!! Especially considering that this advantage is realized every time I do a job. If I do 5 of those jobs a month my 8 core system is making me an extra $10,000 every month!

Now, what was this nonsense about no benefit?

Swing and a miss mate..... As others have pointed out, Feel free to show me the drastic FPS increase of a 8 core over a 2 core to justify the price

 

[Tesselator]

edit - dude, you've just changed the quote to skip the bit about gaming... really rather uncool

Better?

 

[Outsider]

What do you mean? What are the memory support features you're speaking of?

Well the 3500 supports ECC memory (important for some financial and scientific needs). And until we get confirmation from Intel, it's believed that the W3500 series is also not limited to the stock 4.8GBps QPI bus and instead takes advantage of the full 6.4.

 

[Tesselator]

Swing and a miss mate..... As others have pointed out, Feel free to show me the drastic FPS increase of a 8 core over a 2 core to justify the price

How can anyone justify any price difference when discussing $30 games? How about that the machine made me enough money to buy the mostest awesomest gaming box on the planet?? - every month!! No?

Well the 3500 supports ECC memory (important for some financial and scientific needs). And until we get confirmation from Intel, it's believed that the W3500 series is also not limited to the stock 4.8GBps QPI bus and instead takes advantage of the full 6.4.

Hmm, Kewl! Thanks for that. Let's wait and see then. Oughtta be interesting.

 

[nanofrog]

i just love this statement its so true. and kind of not true

brr. if i remember the times when i did a clean reinstall of a windows box. (95,98,xp)
graphics at 800x600 no sound and even keyboard was barely working he he.
so while it does work, it works just barely. i dont know how is it with vista
3 cheers for 3rd party driver manufacturers.

I've had plenty of difficulties of one sort or another in the past with windows.

Vista started out badly, but since SP1, it's made a marked improvement. Even better than my experiences with XP.

And yes, 3rd party drivers are nice. Also quite necessary, as MS depends on them, probably as much, or even more than we do.

The 3500 series is a rebranded Bloomfield with ECC turned on.

Gainestown is markedly different.

Yes, the W35xx is an i7 with ECC. Gainestown isn't as different as you might think. Take a W35xx and add a second QPI.

It's an architectural family, just some features are varied between parts. Classic systems engineering.

I heard a nasty rumor that even the 5500 is essentially the same as the 3500 and i7. That the 3500 and i7 also have a second QPI bus but it is disabled (permanently, so no chance of re-enabling).

From a systems POV, Yes.

What do you mean? What are the memory support features you're speaking of?

ECC