Intel Xeon E5-2670 Deal and Price Tracking

Notice: Page may contain affiliate links for which we may earn a small commission through services like Amazon Affiliates or Skimlinks.

workingnonstop

Active Member
Feb 24, 2016
233
63
28
38
Not sure if it's been done in this thread yet, but some benchmarks I ran -

Dual e5-2670 setup -
Linux-bench: Linux-Bench Linux CPU Benchmarks by ServeTheHome and ServeThe.Biz
Passmark (CPU Mark only): 20k+ CPU mark -> Image of full results

Single e5-2670 in same setup -
Linux-bench: Linux-Bench Linux CPU Benchmarks by ServeTheHome and ServeThe.Biz
Passmark (CPU Mark Only): 13k CPU mark -> Image of full results

Dual e5-2660 in same setup -
Linux-bench: Linux-Bench Linux CPU Benchmarks by ServeTheHome and ServeThe.Biz
Passmark (CPU Mark only): 18.8k CPU mark -> Image of full results

Single e5-2660 in same setup -
Linux-bench: N/A
Passmark (CPU Mark only): 12k CPU mark -> Image of full results
Edited my benchmark post with some fuller results... ran dual/single 2670s and dual/single 2660s in same setup for comparison purposes, if anyone is interested. :)
 

4004

Member
Feb 8, 2016
87
39
18

Attachments

Boddy

Active Member
Oct 25, 2014
772
144
43
In a year or two, will the v2 chips experience the same kind of massive price decrease that the E5 2670 has had?
No need for sarcastic humour and false information when members are asking for advise, not good to steer others down the wrong path. If you want to do comedy, you should state in your post that you 'are just kidding' :D

My understanding is that currently both V1 & V2 CPUs are selling at massive price decrease.
NB. My understanding is that V2 is needed for virtualisation tasks.
 

nthu9280

Well-Known Member
Feb 3, 2016
1,628
498
83
San Antonio, TX
No need for sarcastic humour and false information when members are asking for advise, not good to steer others down the wrong path. If you want to do comedy, you should state in your post that you 'are just kidding' :D

My understanding is that currently both V1 & V2 CPUs are selling at massive price decrease.
NB. My understanding is that V2 is needed for virtualisation tasks.
The SR0KX version (listed by NATEX) E5-2670 supports VT-d. Both SR0H8 & SR0KX support virtualization.
 

Stereodude

Active Member
Feb 21, 2016
454
88
28
USA
My understanding is that currently both V1 & V2 CPUs are selling at massive price decrease.
NB. My understanding is that V2 is needed for virtualisation tasks.
Speaking of false information... v1 E5 are Sandy Bridge-EP. v2 E5 are Ivy Bridge-EP. You're apparently thinking of steppings. v2 E5 Xeons have not yet dropped significantly in price.
 

britinpdx

Active Member
Feb 8, 2013
367
184
43
Portland OR
My understanding is that currently both V1 & V2 CPUs are selling at massive price decrease.
NB. My understanding is that V2 is needed for virtualisation tasks.
I think that it's only the v1 CPU's (Sandy Bridge 32nm) that are massively discounted, but if you have seen v2 (Ivy Bridge 22nm) discounted please let me know where, I'd like to get in on the action.

For the v1 release, the production releases are C1 (SR0H8) and C2 (SR0KX) steppings. Although it has been widely reported on the 'net that C1 steppings do not support VT-d, I (and others on this forum) have posted evidence to the contrary for SR0H8 units, which appear to pass through hardware in ESXI rather well. I'd rather trust my data than internet conjecture ..

BTW, try a google search for "e5-2670 c1 virtualization" and let me know how many of the hits indicate that "C1 stepping is broken" and actually have data that backs it up.

I think there was an Elvis sighting reported on the 'net today, Al Gore really did invent the internet, and Big Al says dogs can't look up ;)
 

Stereodude

Active Member
Feb 21, 2016
454
88
28
USA
Well they didn't revise the silicon and spin a new stepping for no reason. While virtualization may work, or seem to work. Something obviously didn't.
 

Stereodude

Active Member
Feb 21, 2016
454
88
28
USA
  • Like
Reactions: OliG

britinpdx

Active Member
Feb 8, 2013
367
184
43
Portland OR
Thanks for the link. I had searched for a PCN but couldn't find anything, other than than PCN 111178- 00 for the i7 processors that indicated "VT-d erratum removed for C-2 step".
So it's not that vt-d is outright broken on C1, it is operational but a "queued invalidation status write may fail". That explains a lot.
I'm not knowledgeable enough to understand the workaround posted for BT98, nor if was actually implemented by anyone.

What was also interesting to read was the number of errata that carried over from C1 to C2 with "no fix"
 

4004

Member
Feb 8, 2016
87
39
18
For the v1 release, the production releases are C1 (SR0H8) and C2 (SR0KX) steppings. Although it has been widely reported on the 'net that C1 steppings do not support VT-d, I (and others on this forum) have posted evidence to the contrary for SR0H8 units, which appear to pass through hardware in ESXI rather well. I'd rather trust my data than internet conjecture ..

BTW, try a google search for "e5-2670 c1 virtualization" and let me know how many of the hits indicate that "C1 stepping is broken" and actually have data that backs it up.
;)
Do something crazy (see post #29). Like quote yourself.
 
Last edited:

JustinH

Active Member
Jan 21, 2015
124
76
28
48
Singapore
Intel posted a workaround in the Errata for VT-D. I understand a lot of hypervisors implemented the work around (code change) to make it work. (VMWare and KVM confirmed, I imagine Hyper-V as well)

But the workaround has a slight performance hit. Not very noticeable from what I've seen.