I've seen a lot of discussion on these two, but no really definitive answers (including Intel customer support themselves!), so I hope that someone here might be able to provide a little more info...
First...Here is my build:
• Cooler Master Cosmos II case • Corsair Professional AX 1200 Watt Power Supply • Samsung 840 EVO-Series 1TB SSD Drive (for main partition) • 3x WD 4 TB Enterprise Hard Drive(for storage) • Sound Blaster Z PCIe Gaming Sound • ASUS Z9PE-D8 WS Dual LGA 2011 Motherboard • 2x Xeon 2687Wv2 OR 2x Xeon 2697v2 ??? • 64GB CORSAIR Dominator Platinum DDR3 1866 RAM • 3x SLI Evga Titan Black Hydro • 3x ASUS ROG Swift PG278Q (running nVidia surround 2650x1440ea)
I guess the big question is... For both gaming and 3D Rendering... Which will provide the best performance. I know they each rely on different elements, so it would depend on how much the differences between these two chips affect the performance of their various catagories.
Which will have the most impact on my needs? Higher clock speed OR Higher cache, more cores, and lower temps... The other factor is dual 2687Wv2 score higher than dual 2697v2 : http://cpubenchmark.net/cpu.php?cpu=Intel+Xeon+E5-2687W+v2+%40+3.40GHz&id=2045&cpuCount=2
So my real question here is how dramatically does the number of Cores, Cache, and Clock Speed (specifically the differences here) affect Gaming and Rendering?
you will only notice the core difference when rendering - I always tend to stick to the higher ghz unless its server render build - another way to think of it is the 87 is a sprinter while the 97 is in it for the marathon
*its up to you but on builds like this i usually recommend splitting up the budget into two computers or a small render farm on the side so you don't tie down your main rig for 6-24 hours waiting for it to finish rendering - just something to think about, chances are you already have a computer you can use on the side so in that case :p keep on going
Thanks for the great input! I think you are right, but since I already bought these parts, I'm a bit locked in.
My only choice now is the CPU.
Sounds like 2687Wv2 is the way to go for me.... But will the Cache be any factor here with my applications?
At Turbo, the 2687Wv2 has .5GHz more speed than the 2697v2, but 5MB less cache (and 4 less cores as we mentioned, and generates more heat)... is the trade of worth it for .5GHz? How much of an impact will cache have?
The clock-speed advantage is better for games, the extra four cores for the rendering if the software supports them.
I think the deciding factor would be cost; other than the cost of the CPU's themselves is any of the software you are planning to run licensed per core? As already said a second machine to serve as a dedicated gaming system might be a good idea.
Yeah, I think I might consider a second machine down the road. But at this point I'm looking for one rig to rule them all (what everyone wants right?)
So the software I use is:
Autodesk Maya (modeling and rendering with V-Ray and Mental Ray)
Autodesk Mudbox
Adobe Photoshop
Adobe Premiere
And I would like to go back and forth between them. Even better if I can start rendering in Maya, and still work on other software. The balance that I need to strike is that when I do want to play some games, I want to have the best performance I can get on high end graphics intensive games... Crysis, Watchdogs, etc. all maxed out.
So the trick is striking a balance. Here is what I have already purchased for my build.
Cooler Master Cosmos II (single cooling loop to 3x GPUs and 2x CPUs)
Corsair Professional AX 1200 Watt Power Supply
Samsung 840 EVO-Series 1TB SSD Drive (for main partition)
3x WD 4 TB Enterprise Hard Drive(for storage)
Sound Blaster Z PCIe Gaming Sound Card
ASUS Z9PE-D8 WS Dual LGA 2011 Motherboard
64GB CORSAIR Dominator Platinum DDR3 1866 RAM
3x SLI Evga Titan Black Hydro
3x ASUS ROG Swift PG278Q (running nVidia surround 2650x1440ea)
The question is:
2x Xeon 2687Wv2 OR 2x Xeon 2697v2 ??? (or some other dual CPU combo???)
So the big question here is which will yield the MOST performance difference for my purposes
The higher 24 cores over 16 cores are better for rendering, and the higher clock speed 4.0 over 3.5 (in turbo mode) is better for gaming. But I wonder which will be more noticeable to me? Will the 8 extra cores blast through renders 30% faster? Or will the extra .5 GHz give be 10 more FPS while gaming?
I tend to think I should start consider gaming more and lean towards the 2687Wv2 since I'm probably pretty well set on the Maya/Rendering side of things either way. I just want to make sure I don't skimp on my gaming at this point.
Also in benchmark scores dual 2687W v2s outscore the dual 2697v2 ironically.
Well , looking at price , I would get the 16 cores , and keep the extra cash for a 5820k + mobo + whole other pc ( lol ).
The 24 cores would be better in maya , and really , in games you won't see a difference ( probably not ) .
Don't forget , you can raise a bit the base clock to acheive ~2.70ghz on the 24 cores ( did this with my 1231V3 , got it from 3.8 to 3.99ghz with a 105 base clock oc .
Yeah for me, the prices are close enough. Ultimately I want to pick which one will have the best performance for me.
Don't forget , you can raise a bit the base clock to acheive ~2.70ghz on the 24 cores ( did this with my 1231V3 , got it from 3.8 to 3.99ghz with a 105 base clock oc .
This is significant! Although sounds like they would both scale up then.... So can you Turbo AND Overclock with the base clock?
If so, Turbo at 105 base clock on each CPU would be:
2697v2 = 3.7Ghz
2687Wv2 = 4.2Ghz
Is that difference in speed going to make a significant impact to gaming? If not I'll stick with the 2697v2 (already have one, was going to exchange it for the 2687Wv2 if that is much better for gaming, since both should be great for Maya).
Also other than clock speed (which now should be comparable to non-workstation CPUs)... Is there anything else that would make the i7 CPUs better at gaming than the Xeons? If not, then maybe my Xeon build achieves the best of both worlds?
Thank you for this! So according to this you DO Turbo+Base clock, and it stacks. If I'm understanding correctly. And looks like for these Xeons 113 Base Clock will be the magic number.
3.39Ghz with 12 cores running was the number he came up with, but it says it Turbos to 3.5... So will that just happen magically as you are working? And I assume it will go higher than that when the CPU load calls for it?