X79 and/or rampage IV OC'ing thread for those of us still left....

[wanna_buy]

You need different pendrive.
Do you have 16GB one (or less) ?
(micro)SD card with card reader will work too.

I even disassembled my build and unplugged all SATA cables, removed two sticks of RAM and GPU and tried two different 8GB and 32GB drives to no avail. I had an old 4GB drive laying around and when plugged it in Windows, I found out that I had already copied modded X99-A BIOS on it two months ago. How could I have forgotten that! BIOS FlashBack was now successful on my X99 Deluxe II.

I first tried -0.09V undervolt on 103 MHz BCLK, but Windows bluescreened when I was unzipping Cinebench R15. I then applied -0.05V undervolt. The CPU frequency is 3500MHz when idle, 3400 MHz when running Cinebench R15. However, the frequency is barely 3 GHz when running Cinebench R20. I tried 105 MHz BCLK, the system seems to boot, but there is no video signal and the screen just remains black.

Behold my Cinebench R15 score. It is even higher than it was on X99-A on the BIOS same settings.


Now where is bobbybluz with his promised E5-2699V4 score?

 

[bobbybluz]

I even disassembled my build and unplugged all SATA cables, removed two sticks of RAM and GPU and tried two different 8GB and 32GB drives to no avail. I had an old 4GB drive laying around and when plugged it in Windows, I found out that I had already copied modded X99-A BIOS on it two months ago. How could I have forgotten that! BIOS FlashBack was now successful on my X99 Deluxe II.

I first tried -0.09V undervolt on 103 MHz BCLK, but Windows bluescreened when I was unzipping Cinebench R15. I then applied -0.05V undervolt. The CPU frequency is 3500MHz when idle, 3400 MHz when running Cinebench R15. However, the frequency is barely 3 GHz when running Cinebench R20. I tried 105 MHz BCLK, the system seems to boot, but there is no video signal and the screen just remains black.

Behold my Cinebench R15 score. It is even higher than it was on X99-A on the BIOS same settings.
View attachment 260976

Now where is bobbybluz with his promised E5-2699V4 score?

I ran into some more serious medical issues (ended up in the hospital, see a kidney specialist next week) and working on my 2699 V4 rig has been near the bottom of my priority list. I'll eventually get around to it but at the moment it's sitting parked with nothing connected to it. I have a loaned 6950X to play with in it as well but that's gathering dust too. I've been working on my two Z690 Alder Lake builds as energy permits and after installing the Thermalright LGA 1700 anti-bending CPU frames on both was finally able to get Cinebench R20 to run on the 12700K @5.25GHz without crashing last night. The 12600K rig isn't finished far enough to run any benchmarks on it yet and with Win 11 Pro 22H2 due out on the 20th is awaiting it before proceeding any further on that project. The 12700K has Win 11 Pro 21H2 on it and will likely stay that way for the time being.

Then there's the defective Arctic Liquid Freezer II cold plate debacle and I have three that are affected. I have the 240 that was in the 12600K out and replaced it with an unaffected 280 that was in the 12700K build that'd been replaced with a new 420. I still have to pull the pumps on my X99 E5 2683 V4 and X79 E5 1680 V2 so I can photograph the production codes as required by Arctic in order to get the free upgrade/repair kits. It seems like a really simple thing to do but when I run out of energy drastically fast at the moment it turns into a real project. Things I used to be able to do in a few minutes sometimes takes me a few days now and it's no fun at all.

 

[wanna_buy]

Can BCLK overclocking with too many SATA devices attached cause random lockups or data corruption? I have 1 SATA SSD and 2 SATA HDDs installed in 18-core Xeon system.

When I first launched Fortnite, the system froze on the loading screen. I then played for 1 hour without hiccup. Upon the restart, the system froze during web-browsing. There were no errors in Event Viewer. I reduced undervolt to -0.04V and ran Prime95 overnight. When I saved the screenshot of HWinfo temps and opened the file while running Prime95, the photo viewer screen was blank. I stopped Prime95 and reopened the screenshot and I got Windows error saying File System error. There were no Corrected Hardware Errors in Event Viewer. After I gamed for a while, the system locked up randomly during web-browsing. I take it that BCLK overclocking causes instability, so I dropped BCLK to default 100, but left undervolt and see what happens next. The multiplier during Cinebench R15 run is still 33x.

I still got random lockup on 100 BCLK. I reverted overclocking and restored everything to Auto. If my system again freezes randomly, then the culprit must be the HDD.

 

[wanna_buy]

I was running X99-Deluxe II system at stock for 10 days without any random lock-up. Possibly that Xeon chip could not take undervolt. Yesterday I swapped out the CPU for third E5-2699V3 I had laying around. I got 2977 scores in Cinebench R15 with -0.07 undervolt, but soon Windows blue screened. When I restarted, the saved screenshot file was corrupted and could not be opened. I even tried -0.02V undervolt on this third E5-2699V3 chip for complete stability, but still got random lockup during web browsing after 2 hours. I began to suspect that something else in my current setup (iffy motherboard or 4x16GB RAM) was causing instability.

I installed second Xeon on Rampage V Extreme with 4x8GB RAM. With -0.07 undervolt, the 3000 points barrier in Cinebench R15 was broken. I have beaten johnspack's record with 103 BCLK and 3500 MHz frequency.


Interestingly enough, my first E5-2699V3 (which I sold 2 months ago) on ASUS X99-A scored lower at the same 3500 MHz frequency. Any explanation?


I went on a rampage and gained #1 spot in nine categories for E5-2699V3 on hwbot. 105 BCLK was benchmark-stable on Rampage V Extreme. However, such high BCLK after the first boot and restart resulted in no boot before clearing the CMOS via the dedicated button on the back. 105.5 BCLK could boot into Windows, but the benchmarks crashed. 106 BCLK resulted in no start. 103 BCLK is the real threshold for 24/7 stability.

 

[bobbybluz]

Cinebench R15 and R20.

 

[wanna_buy]

Cinebench R15 and R20.

Did you overclock your CPU? What are the frequencies under load?

 

[bobbybluz]

Did you overclock your CPU? What are the frequencies under load?

You can't OC a V4 Xeon. They're locked as far as multiplier and RAM speed. You can play with the BLCK a bit but that's about it. The V3 turbo boost trick doesn't work on them either. The GPU in that rig is ancient too. The OS is Win 10 LTSC 2019. I hadn't powered it up in a few months until tonight.

 

[wanna_buy]

You can't OC a V4 Xeon. They're locked as far as multiplier and RAM speed. You can play with the BLCK a bit but that's about it. The V3 turbo boost trick doesn't work on them either. The GPU in that rig is ancient too. The OS is Win 10 LTSC 2019. I hadn't powered it up in a few months until tonight.

To be honest, I never expected such high scores on stock settings. I assumed since my stock E5-2699V3 posted 2296 scores and V4 had 22% higher core count, it would score about 22% higher or around 2800. This guy got around 2640 scores on his E5-2699V4.

Your CPU surpasses even stock Ryzen 3900X. Here is the screenshot of my Cinebench run on Ryzen 9 3900X.

 

[wanna_buy]

I have achieved #1 spot in 18 categories for E5-2699V3 on hwbot. https://hwbot.org/hardware/processor/xeon_e5_2699_v3/

GPUPI 3.3 would not even run on my system, throwing some error about running out of the resources. GPUPI 2.1 only runs on NVIDIA cards. I ran PCMark Vantage 2 times, but it did not give me any score after running for more 1 hour each time.

I got more scores in Geekbench 3 single core test than the current #1 spot holder for the multi score. How is that even possible that he scored higher on the lower CPU frequency?

I also borrowed 64GB RAM from another system to be able to run y-cruncher - Pi-10b. I run the benchmark 2 times, but the test took 489 seconds.

 

[VeqIR]

I've just updated the system with an NVMe drive finally (2TB Samsung via a PCIe adapter), and I want to get back to sorting out RAM settings.

EDIT: I've figured it out -- it was actual RAM stick issue that caused the below problems.

A year ago I swapped from an i7 to a Xeon E5-1680V2. During the installation I bent a couple of socket pins, which had never happened to me before in lots of installs over the years, so I thought the system was just doomed. I carefully bent the pins back to about where it seemed the tips should be located, and miraculously the system booted fine. For the past year everything's been solid. Except I just can't do any overclocking with RAM, and the processor itself doesn't seem stable with overclocking. So I don't know if I got a bad processor (it was used, pulled from some workstation), or if the bent socket pins don't quite make the right contact after all and are causing this issue. Logically speaking, I feel like if the pins were at fault, I'd see instability at stock settings too, but I'm not. It's been fine for the past year and a few months, through gaming and general use, just at completely stock settings =/

There are 32GB of RAM in the system in 4 sticks. With an i7 4930K that used to be on the board before the Xeon, I had the processor overclocked a bit to 4.4 GHz per core, and the RAM was in XMP1 profile with no manual voltage bumps, just keeping everything on auto and with a slight negative offset, and all the temperatures and voltages were good, well within safe limits. But with the Xeon, XMP1 profile results in WHEA blue screen errors. XMP1 is not stable. I've tried running gradual overclocking and running MEMTEST86 from a USB stick for stability testing after each change, and invariably I'd eventually get a WHEA error blue screen, until I went back to stock. Then it runs fine without errors.

So. What do you think: bad processor or the damaged socket pins at fault? Are these Xeons extra picky with RAM settings? Maybe this processor type does not work properly with XMP profiles or 4 RAM sticks in XMP? Is there some particular voltage I should be bumping up perhaps to help?

(motherboard is ASUS Rampage IV Black Edition on a more recent modded BIOS that had meltdown and spectre patches and enabled NVMe).

 

[Zyll Goliat]

I've just updated the system with an NVMe drive finally (2TB Samsung via a PCIe adapter), and I want to get back to sorting out RAM settings.

A year ago I swapped from an i7 to a Xeon E5-1680V2. During the installation I bent a couple of socket pins, which had never happened to me before in lots of installs over the years, so I thought the system was just doomed. I carefully bent the pins back to about where it seemed the tips should be located, and miraculously the system booted fine. For the past year everything's been solid. Except I just can't do any overclocking with RAM, and the processor itself doesn't seem stable with overclocking. So I don't know if I got a bad processor (it was used, pulled from some workstation), or if the bent socket pins don't quite make the right contact after all and are causing this issue. Logically speaking, I feel like if the pins were at fault, I'd see instability at stock settings too, but I'm not. It's been fine for the past year and a few months, through gaming and general use, just at completely stock settings =/

There are 32GB of RAM in the system in 4 sticks. With an i7 4930K that used to be on the board before the Xeon, I had the processor overclocked a bit to 4.4 GHz per core, and the RAM was in XMP1 profile with no manual voltage bumps, just keeping everything on auto and with a slight negative offset, and all the temperatures and voltages were good, well within safe limits. But with the Xeon, XMP1 profile results in WHEA blue screen errors. XMP1 is not stable. I've tried running gradual overclocking and running MEMTEST86 from a USB stick for stability testing after each change, and invariably I'd eventually get a WHEA error blue screen, until I went back to stock. Then it runs fine without errors.

So. What do you think: bad processor or the damaged socket pins at fault? Are these Xeons extra picky with RAM settings? Maybe this processor type does not work properly with XMP profiles or 4 RAM sticks in XMP? Is there some particular voltage I should be bumping up perhaps to help?

(motherboard is ASUS Rampage IV Black Edition on a more recent modded BIOS that had meltdown and spectre patches and enabled NVMe).

What is the RAM speed?If the RAM speed is 2133Mhz or higher maybe you should try to put it down a bit and see how that goes.......sometimes Xeons could be tricky when you OC them and if you trying to run them on higher memory speeds.....anyway you should try manual memory tweaking....

 

[VeqIR]

I think I actually did mess it up with the RAM channel connections... It's only reading 3 sticks of RAM now instead of the 4. I guess after all it's time for an upgrade and an honorable retirement for this board

 

[agent_x007]

If you push frequency too high (or timings too low), you will drop memory channels.

 

[lexluthermiester]

I think I actually did mess it up with the RAM channel connections... It's only reading 3 sticks of RAM now instead of the 4. I guess after all it's time for an upgrade and an honorable retirement for this board

If the RAM works fine at stock clocks but goes wonky with an OC, your system is fine and your RAM is not overclockable. RAM overclocks make very minor improvements, so it's not worth bothering to do.

 

[VeqIR]

Thanks all, I've figured it out, it was a RAM stick problem. A particular stick went bad and was causing overall instability. It used to work properly for a few years but something must have happened to it over time. It wasn't overclocked per-ce, but it was set to XMP1. Maybe it didn't like that voltage. It finally stopped being read at all after I took it out for cleaning yesterday. It's registers in HWInfo64 as being present in slot D1, but noted differently (by its full product name, rather than speed and timings), but it's not read by the BIOS or Windows as contributing to the total RAM amount. Living with 3 sticks now; 24GB of RAM is still okay (not 4 channel now though =/) Just one more step closer to upgrading to a modern system. Hopefully after the 13th Gen of Intel processors are out.

 

[lexluthermiester]

Living with 3 sticks now; 24GB of RAM is still okay

Get a replacement stick. Shouldn't be too difficult to find.

 

[1000Gbps]

Hi again, returned to my hobby of trying to find what's killing my X79 system (random crashes) and found something strange.
The current CPU E5-2689 is stable only on 2.3GHz per core, at 2.4 it crashes after 15-20min, when the system is in auto it's trying to boot then crashes during Windows boot screen.
Motherboard is AsRock X79 Extreme 9 BIOS ver. P4.0, populated with 8 sticks of Kingston 1600@stock freq (also auto mode). This is the latest CPU used on this mobo, before that there were E5-2660, E5-2650, all used on 2.3-2.4GHz and always crashed at auto mode (that's why I bought more E5s, thought they are somewhat defective)

Any ideas?

 

[VeqIR]

I want to add that I've found that my ASUS RIVBE board is highly sensitive to any flexing, as it's a very large board overall. If you have a heavy heatsink on the CPU, it can lead to all sorts of instabilities. I've been using an MSI GPU riser post to help prop up the CPU heatsink (BeQuiet! Dark Rock Pro 4 in my case) and the GPU for a good measure too. I've never had this issue with other builds thus far, but at least twice it's been a problem with the RIVBE board.

 

[1000Gbps]

CNPS10X, it's not heavy to bend the mobo and the CPU is very cool (this summer 54°C max because of these sh*tty 2.3GHz)

 

[drizzler]

Just did some benchruns with my 1680v2 before swapping to x99, 1.375v under load for 4758ghz is ok but coming from 1.225v for 4.5ghz it was not worth it in the past.

3DMark.com search

3DMark.com search

www.3dmark.com

Intel Xeon E5 1680 v2 @ 4757.65 MHz - CPU-Z VALIDATOR

[i3zl63] Validated Dump by DESKTOP-TN40AH1 (2022-09-28 17:08:05) - MB: Asus RAMPAGE IV BLACK EDITION - RAM: 16384 MB

valid.x86.fr

 

[wanna_buy]

I now hold #1 spot in 21 categories for E5-2699V3 on hwbot.

Just did some benchruns with my 1680v2 before swapping to x99, 1.375v under load for 4758ghz is ok but coming from 1.225v for 4.5ghz it was not worth it in the past.

3DMark.com search

3DMark.com search

www.3dmark.com

View attachment 263455

View attachment 263484

Intel Xeon E5 1680 v2 @ 4757.65 MHz - CPU-Z VALIDATOR

[i3zl63] Validated Dump by DESKTOP-TN40AH1 (2022-09-28 17:08:05) - MB: Asus RAMPAGE IV BLACK EDITION - RAM: 16384 MB

valid.x86.fr

You should install Benchmate and run Cinebench R23. Then save results with both CPU-Z CPU and Memory Tabs open. Your score is higher than the current #1 spot holder on hwbot.

I have improved my own records on hwbot and currently hold #1 spot in 21 categories for E5-2699V3.

 

[r3down]

Hi guys I might be a little late to the party here-
Im on a 3930k w 32gb 1866 ram, on an asus RIVE
and considering buying E5-2697v2 or greater for an upgrade
but ive never done bios injection before, any input is appreciated.

 

[wanna_buy]

Hi guys I might be a little late to the party here-
Im on a 3930k w 32gb 1866 ram, on an asus RIVE
and considering buying E5-2697v2 or greater for an upgrade
but ive never done bios injection before, any input is appreciated.

Xeon BIOS injection does not work on X79 motherboards, only on X99.

I was scrutinizing HWBot submissions and found out that Gigabyte GA-X99-SOC Champion is capable of reaching 109.3 BCLK.

sofos1990`s Geekbench3 - Multi Core score: 48209 points with a Xeon E5 2699 v3

The Xeon E5 2699 v3scores getScoreFormatted in the Geekbench3 - Multi Core benchmark. sofos1990ranks #71 worldwide and #1 in the hardware class. Find out more at HWBOT.

hwbot.org

This guy even managed to boot into Windows with 106.7 BCLK on Rampage V Extreme.

NAMEGT`s wPrime - 32m score: 2sec 533ms with a Xeon E5 2699 v3

The Xeon E5 2699 v3 @ 2987.7MHzscores getScoreFormatted in the wPrime - 32m benchmark. NAMEGTranks #86 worldwide and #2 in the hardware class. Find out more at HWBOT.

hwbot.org

 

[r3down]

It could be how I started explaining it-

asus has a xeon 2697v2 under its supported cpus-

so i started to go on this journey,

Perhaps a better question could be-

what options are out there at the moment?

 

[bobbybluz]

For X79 E5 1680 V2's are unlocked 8 core Xeons. I have two of them. Their performance is nearly identical to a X99 5960X when both are overclocked to the same speed (4.5GHz for mine). The 2697 V2's have four more cores but are also locked.