Overclock.net banner
1,461 - 1,480 of 1,897 Posts
Quote:
Originally Posted by ucode View Post

@tbob22 2680 v0 or v2? Was that 3.3GHz 31x multi plus bclk OC or at 33x multi?
Sandy Bridge/v0, the video says it was 3.3ghz on all cores for the e5-2680, he didn't mention any bclk OC, so I'm guessing that was stock.

On my E5-2670 I'm pretty sure it was 31x on all cores without bclk, because with bclk OC was right at 3.3ghz and I was limited to 106mhz if I remember right.

Edit: Found cinebench result:


Was running single channel during testing, so score is a little low.
 
Quote:
Originally Posted by tbob22 View Post

Sandy Bridge/v0, the video says it was 3.3ghz on all cores for the e5-2680, he didn't mention any bclk OC, so I'm guessing that was stock.

On my E5-2670 I'm pretty sure it was 31x on all cores without bclk, because with bclk OC was right at 3.3ghz and I was limited to 106mhz if I remember right.

Edit: Found cinebench result:


Was running single channel during testing, so score is a little low.
u see i have a e5-2670 v0 with rampage 4 BE (x79) .
See turbo details of e5-2670 here>>
http://www.cpu-world.com/CPUs/Xeon/Intel-Xeon%20E5-2670.html

i can run all core @3.0Ghz with 100bclk , which is default turbo.
So u saying it is possible for me to get 3.3ghz @all core with these v0 like e5-v3s on x99 ?
 
Quote:
Originally Posted by mohiuddin View Post

u see i have a e5-2670 v0 with rampage 4 BE (x79) .
See turbo details of e5-2670 here>>
http://www.cpu-world.com/CPUs/Xeon/Intel-Xeon%20E5-2670.html

i can run all core @3.0Ghz with 100bclk , which is default turbo.
So u saying it is possible for me to get 3.3ghz @all core with these v0 like e5-v3s on x99 ?
No, only with the 2680 according to the video. 2670 should get 3.1ghz @ 100bclk if you adjust the short and long power duration limits so it can stay in turbo longer.
 
Discussion starter · #1,465 ·
Hi guys, glad to see the thread getting a boost
wink.gif


I'm actually not the best person to ask as my cunning plan for an upgrade to x79 was stymied by funds being unavailable.

To tbob22 I am indebted, as ever.

Wendell can surely be trusted too
smile.gif
 
Maybe you guys will dig this, maybe not. Had an idea to try to take an HP Z620/Z420 motherboard that i thought was X79 (its C602, which is basically the same still) and repurpose it into a regular desktop for gaming and whatever else like i did with an HP Z400 board. I solved all the sense pin issues with the Z400 board and had it in a standard atx case using a standard atx power supply with an adapter i made. The Z620/420 boards have PSU adapters readily available, but the sense pin issues are worse and i couldn't solve it. Ended up scrapping the idea of using it for cheap xeon budget gaming builds for people, but I decided since i had the motherboard I can use it on my newly made test bench. Only issue is it says "press F1 to boot" every time you turn it on, which isn't much of an issue for me. The Z400 one i figured out, that boots thru straight to windows with no issues now.

Anyways, i found an E5-1620 for $40 and threw it in there with 8gb of RAM and it works great! I can swap GPUs and other parts as needed, temps are fine with a hyper 212 evo i had lying around, and i occasionally use it as my streaming PC for encoding. I know that with a proper x79 motherboard i can OC this CPU, but for what im using it for this way its fine.

I pair that motherboard with an E5-1650 that later found life in an EVGA x79 SLI board i bought for $100 with bent pins and had repaired (eventually got that 1650 up to 4.5ghz which was cool) in this vid:
and detail the test bench setup it ended up at with the E5-1620 in this one.
 
Quote:
Originally Posted by ucode View Post

Those with Asus X79 and v2 could try here for high all core turbo. Would be interesting to see the results.
Thanks. But any proof that a v2 xeon is running at its full glory?
A 2680 v2 all core @3.6ghz at bclk 100 would be awesome
Come onnn... Guys.
Make some noise in this thread.
 
Quote:
Originally Posted by mohiuddin View Post

Thanks. But any proof that a v2 xeon is running at its full glory?
A 2680 v2 all core @3.6ghz at bclk 100 would be awesome
Come onnn... Guys.
Make some noise in this thread.
I can´t find where he has the bioses. I posted in the thread at Anandtech that I would like him to share the bioses. My 2680v2 works with 113Mhz bclk so that would be 4Ghz on all cores
drool.gif


But I guess TDP at 115w will not accept that. Unless TDP can be modified as well.

EDIT: When I load all cores now at 3.5Ghz, HWMonitor reports that it uses 105w so there is some headroom. The 2690v2 has a TDP of 130w.
 
So i feel like i've been having temp issues, or higher running temps on my brother's E5-1650 (v1). I know it runs hot but something seems off. Motherboard is an EVGA X79 SLI, has 4 sticks of 4gb DDR3 1600 Corsair Vengeance LP @ 1600mhz.

Originally i was able to get to 4.5ghz @ 1.394v with max stress testing using p95 peaking at 85-87c. Vdroop configured as "disabled." I figured this was okay, most stuff wont stress the CPU that hard. Eventually the bad airflow of this case (Raidmax Delta, think like Phanteks P400) eventually caused this to run hotter and hotter once i gave the system to my brother. I brought the OC down to 4.2ghz @ 1.280v and this seemed to peak at 82c in the same environment. Then the pump on this Arctic Liquid Cooler 240 died, and i threw on a Hyper 212 evo which i had to run the CPU at stock mutli with 1.2v vcore to stay manageable. Then in that same case, the EK Fluid Gaming 240 CPU only kit was added and i got the OC back up 4.5ghz at 1.394v at better temps, but eventually this too deteriorated. My brother had to resort to using his PC without the front panel on (rad was on front intake). This made his rad get super dusty super quick and things only got worse. Eventually, it appeared the plastcizer on the tubing of the a240 kit leeched into the block and i had to put the CPU back to stock settings to run acceptably (still without front panel).

Then my brother got an Inwin 303, mayhems tubing, and redid the loop and got the block and stuff cleaned out and its back to normal. Here is the system:



There are 3 fractal venturi fans on the bottom, 2 of which the big ass Gigabyte Xtreme Gaming 980ti gets in the way of, and 1 of which the DDC pump/res is mounted on top. So i imagine airflow upwards is sort of...not at its best. The radiator is mounted up top behind those EK Vardar fans and the rear fan, originally set up as exhaust, has been flipped to intake.

Now this time, 4.5ghz at 1.394 was too hot. I had to adjust the voltage down, and bring the OC down to 4.3ghz to get mid-to-high 80s C in p95 smallFFT (15 mins). HWmonitor reported 215w of power usage on the CPU. Left it there since i ran outta time, but seemed okay. Initial streaming (OBS x264 encoding) + gaming had it stable at mid-60s C not breaking 70. My brother reported the occasional crash to desktop/blackscreen/etc so i bumped the voltage up a notch. Then over the weekend he said he was running 82c across all cores while gaming + streaming and wanted the OC backed off again. Actually he demanded it back at stock, so i set it back and ran p95 SmallFFT again. 80c peak, no OC and stock/auto voltage. Okay now, what the hell? Reset bios to optimized defaults and started over and now temps were high 60s/low 70s at stock under p95 smallFFT, but HWmonitor reporting 180w of power usage. I noticed vdroop went from "disabled" to "intel SPEC" when i reset the BIOS. Left it on auto, upped the multi to 40 for 4ghz and hit p95 smallFFT again. Reported power usage of 194w, temps from high 60s to mid-70s. Ran it for like 30 mins while browsing and doing other stuff to stress it out. Seemed solid as a rock. Gaming didn't break 60c. He hasn't had time to try streaming + gaming again, but it shouldnt be higher than p95 smallFFT (well thats what i told myself before but he still claimed 82c across all cores which was maybe 3-5c off p95 smallFFT at that point in time). So the intel spec vdroop setting with voltage on Auto and multi at 40 has it pushing 1.31v (according to HWmonitor).

Now im thinking it runs everything great at 4.0 and my brother doesn't want to deal with crazy temps anymore. How should i approach this now to bring temps down? Should i manually set voltage to 1.2v and see how it responds? This has been a weird ride with this Xeon.
 
Quote:
Originally Posted by Helgaiden View Post

So i feel like i've been having temp issues, or higher running temps on my brother's E5-1650 (v1). I know it runs hot but something seems off. Motherboard is an EVGA X79 SLI, has 4 sticks of 4gb DDR3 1600 Corsair Vengeance LP @ 1600mhz.

<<snip>>
To start, 1.394v for 4.5ghz seems really high on this chip. How did you get to that voltage? I've had four 1650's and they could all do 4.5ghz at 1.3v or less. My 1660 @ 4.6ghz maxes out around 210w SmallFFT's w/AVX according to hwinfo64, I suppose it could be that evga board or just a poor clocker.

How are all the other voltages looking? VTT, VCC PLL, etc If they aren't all manually set, I'd start there if anything is on auto the board may be pushing VTT way up or something.

Temps seem really high as well, with my D15 it maxes around 70c after a few hours, but voltages are also lower.
 
Quote:
Originally Posted by Korgan View Post

I can´t find where he has the bioses. I posted in the thread at Anandtech that I would like him to share the bioses. My 2680v2 works with 113Mhz bclk so that would be 4Ghz on all cores
drool.gif


But I guess TDP at 115w will not accept that. Unless TDP can be modified as well.

EDIT: When I load all cores now at 3.5Ghz, HWMonitor reports that it uses 105w so there is some headroom. The 2690v2 has a TDP of 130w.
Please let us know when he shares any.
And isn't there a to change the tdp in bios? I m using rampage 4 BE by the way
 
Quote:
Originally Posted by tbob22 View Post

To start, 1.394v for 4.5ghz seems really high on this chip. How did you get to that voltage? I've had four 1650's and they could all do 4.5ghz at 1.3v or less. My 1660 @ 4.6ghz maxes out around 210w SmallFFT's w/AVX according to hwinfo64, I suppose it could be that evga board or just a poor clocker.

How are all the other voltages looking? VTT, VCC PLL, etc If they aren't all manually set, I'd start there if anything is on auto the board may be pushing VTT way up or something.

Temps seem really high as well, with my D15 it maxes around 70c after a few hours, but voltages are also lower.
Heres a pic with the voltages, but before i reset the bios. Though everything should be mostly the same as a i never touched anything other than Multi, vcore, and vdroop.



Even when fed 1.394, prime95 would drop a worker occasionally (not always) during long tests, so i guess i ended up at that voltage for two reasons: 1- aforementioned p95 issue. 2- originally i set the voltage in the 1.25 range to start but i noticed with vdroop to disabled, the board added like 0.11v so i had to factor that when looking for something that appeared stable.
 
-
Quote:
Originally Posted by Helgaiden View Post

Heres a pic with the voltages, but before i reset the bios. Though everything should be mostly the same as a i never touched anything other than Multi, vcore, and vdroop.



Even when fed 1.394, prime95 would drop a worker occasionally (not always) during long tests, so i guess i ended up at that voltage for two reasons: 1- aforementioned p95 issue. 2- originally i set the voltage in the 1.25 range to start but i noticed with vdroop to disabled, the board added like 0.11v so i had to factor that when looking for something that appeared stable.
Those VSA/VCCSA and VCCIO/VTT voltages are way too high. I have mine at 1.1v. 1.2v is the max I'd ever recommend, any higher and your IMC could fail.

PLL should be 1.8v, There shouldn't be a reason to go above that. Other voltages look OK, but I'd manually set them anyway to be safe.

A bit of info on auto voltages and why they should never be used when overclocking:
https://rog.asus.com/forum/showthread.php?19926-3930K-C2-Degradation
 
Quote:
Originally Posted by tbob22 View Post

-
Those VSA/VCCSA and VCCIO/VTT voltages are way too high. I have mine at 1.1v. 1.2v is the max I'd ever recommend, any higher and your IMC could fail.

PLL should be 1.8v, There shouldn't be a reason to go above that. Other voltages look OK, but I'd manually set them anyway to be safe.

A bit of info on auto voltages and why they should never be used when overclocking:
https://rog.asus.com/forum/showthread.php?19926-3930K-C2-Degradation
Oh. Well um, do you think degradation is now an issue as a result?

Also, what are all the voltages i should manually set then?
PLL -> 1.8v
VSA -> 1.1v?
VCCIO -> 1.1v?
 
Quote:
Originally Posted by Helgaiden View Post

Oh. Well um, do you think degradation is now an issue as a result?

Also, what are all the voltages i should manually set then?
PLL -> 1.8v
VSA -> 1.1v?
VCCIO -> 1.1v?
It's possible.. I'd start with those settings and see how it goes. The high temps could definitely be due to the auto voltages.
 
  • Rep+
Reactions: Helgaiden
Quote:
Originally Posted by Helgaiden View Post

Cool ill swing by to try those with him today!

There is also this page i took a pic of if there is any other advice!

The default power settings are usually fine.

EIST and C1E are disabled, I usually leave them enabled to get some power savings when idle. But for testing I guess having that all disabled is OK.
 
Hey guys I tried reading through the pages to see if anyone had input on the E3-1200 series up to 1275 in a LGA1155 mobo, not to get too off topic but did anyone ever comment if it worked well on a LGA1155 board. Was helping a good friend of mine and made the rec that he should go that route vs the i7 but now I was not sure if I was right since if all prices are the same I always found the xeons to perform real well
 
Snagged another x79 Deluxe for $150, gonna drop a e5-2670 in it.
smile.gif
 
1,461 - 1,480 of 1,897 Posts