• Competitor rules

    Please remember that any mention of competitors, hinting at competitors or offering to provide details of competitors will result in an account suspension. The full rules can be found under the 'Terms and Rules' link in the bottom right corner of your screen. Just don't mention competitors in any way, shape or form and you'll be OK.

1366 X58 Xeon 5650

Anyone got any more benchmarks? Also, is the X5660 the favored variant of the Xeon?

Really tempted just to go for it.

http://www.xbitlabs.com/articles/cpu/display/core-i7-2600k-990x.html

That may help. Obviously in most tests the 2600k will win (it's 9.6% faster in lightly threaded apps) but once the support is there the 990x wins quite easily.

As our tests showed, the recently launched Core i7-990X Extreme Edition processor with maximally increased clock speed can only compete against Core i7-2600K in a few individual cases – in those few applications that create heavy load that could be well-paralleled.

However you need to remember that the article was written over three years ago, and we now have that support far more than we did back then.

IE - overall when it's all balls to the wall the Xeon will be a faster CPU than the 2600k. I would imagine that's a pretty huge upgrade if you are used to using a 920 or 950.

The Xeon is 32nm, the 9x0 is 45nm. It's a completely different technology.

TBH the only reason the Xeons fell short of Sandybridge on overall perfomance was that they don't overclock anywhere near as far. Sandy 2600ks mostly did 5ghz (or less once people realised it was a bad idea to shove them full of volts) whereas the Xeons need to be really well binned to reach 4.5ghz on water.

Either way the Xeon (or 980/990X) is a formidable upgrade from a quad core I7 on X58.

but they're both great CPU's and compare very well to the 4770k @ 4.4Ghz.

Yeah TBH I would say that 100% utilised that's about a very fair assumption. BTW as promised.. Here is my 3970x @ 4.7ghz.



So I'm only around what? 15% higher than your results. I can push on to 4.8ghz but then I run close to the edge...

Cine is one of the best apps for testing a CPU's overall might.
 
Last edited:
Thanks for the heads up chaps. Going to go for a 5650 and hopefully will get it to 4Ghz :)

Edit: bagged a X5650 for £50 inc delivery. Worth a short. Worse case scenario I'll keep it until I upgrade to X99 and use it in my current board as a home lab with 12 threads. Yummy.
 
Last edited:
I've also gone for a X5650 having read through this thread. Should be a good upgrade from my very hot/poor clocking 920 C0.

If I've read it correctly, the i7 920 is 135w where as the Xeons are 95w - that's got to be better too.

Excited!
 
I've also gone for a X5650 having read through this thread. Should be a good upgrade from my very hot/poor clocking 920 C0.

If I've read it correctly, the i7 920 is 135w where as the Xeons are 95w - that's got to be better too.

Excited!

My 920 is also a C0 that won't clock past 3.4 unless I give it a obscene amount of volts.

Must admit I didn't think I would get this excited over a £50 processor. Shame I probably won't get everything (also needed some new thermal paste, etc) before the weekend :(.
 
For the price (~£50) I thought it was worth a shot, currently got my 920 C0 at 3.8ghz stable with 24gb ram (no voltage change) so if the X5650 can match it I'll be happy.
 
Thinking ahead, what sort of coolers do these require? I currently have a Noctua NH-U12 which has been in since I built the system. Will this be ok?
 
Anybody know if the Asus Sabertooth will take the Xeon's? There is no listing on the Asus website and the last bios update mentions nothing about cpu updates.
 
Is there any performance benchmarks of actual games vs newer processors? I want to know if the minimum fps is affected much or if there isn't much in it providing using the same gpu, all i can find are synthetic test scores..
 
Anyone tried the x5680 or x5690 or does the purchase price outweigh any tangible benefit on x58 for gaming? I can see some x5680 for around £230-240 from the US and have 25 multiplier! The x5690 has a 26 multiplier but costs considerably more.

From what I've seen the multipliers are as follows:

x5650 - 20
X5660 - 21
X5670 - 22
X5675 - 23
X5680 - 25
X5690 - 26

But is that only part of the story with these chips? What's the general consensus for the preferred chip to tide i7 920/930 users over until skylake? I think I will try and snag an x5675 unless people suggest a reason not to...
 
I was thinking the same thing of either a 5670/5675

From the reading i did yesterday it seems that some X58 boards struggle with BLCK of over 200, so getting something with a higher multi should make it easier to get 4GHz+
 
I've upated my benches to include the 5670 that I bought and clocked to 4.4Ghz. Performance wise, its not a great leap over the 5650 @ 4GHz and I'm not sure it justifies the extra money but they're both great CPU's and compare very well to the 4770k @ 4.4Ghz.

bench3.jpg


bench4.jpg

yeah thanks for posting..helped me in my decision.
 
No probs, I've got my ram running at 1600MHz now and its improved Cinebench to 987 but not made any real difference to the other scores.
 
I may have jumped the gun as I'm not sure my mobo can support Xeon processors though I thought I saw some people in this thread with the same mobo and Xeon CPU. Just tried googling the asus website and it doesn't show that the p6x58d-e is compatible. Has anyone got a definitive answer?
 
I may have jumped the gun as I'm not sure my mobo can support Xeon processors though I thought I saw some people in this thread with the same mobo and Xeon CPU. Just tried googling the asus website and it doesn't show that the p6x58d-e is compatible. Has anyone got a definitive answer?

Its on the compatible list from post 42 of this thread. Looks like a great board to use with it having modern features like USB3 and SATA3
 

Sorry to dig up an older post but this isn't how you use IBT. "Standard" isn't a stress test.

Without getting into the details, you need to increase the memory usage so that the GFLOPS column is higher, at least 80 at that speed with 6 cores.

It will take longer per loop but will actually be a good stress test then.

Only problem with IBT is that it doesn't come with instructions.

Edit: corrected because it's Gulftown and doesn't have AVX, so GFLOPS = 4 * cores * clock speed, not 8 as it is for SB/IB.
 
Last edited:
Back
Top Bottom