Xeon Owners Club

[frankr2994]

Actually playing with dmi clock speed and dmi/peg ratio right now if anyone has any input. Testing on a 2630v3. Moving away from the turbo hack to maybe get some better performance. Should work on the v4s as well. Once figured out I'm going to buy a pair of better CPUs.

 

[CapitanXeon]

Everything i tried when i installed the CPU resulted in a no boot, but i also have 8 ram sticks and i suspect the memory controller in this isn't all that happy, but so far no problems at everything stock speeds.

 

[Edwired]

It works fine on 10 for sure, but the PreBoot migration will bsod every time with inaccessible boot volume when running the 11 installer, same deal if you install them after, i had to dig through some old notes i had about where windows references the pci device to the appropriate driver to revert it too, as the installation failed to create a restore point (and windows no longer implements "Use last known good configuration")

Unless you didn't configured the setup and other files to compatibility mode for it to work on windows 11 what motherboard you trying to use?

 

[SiggiTheHun]

Good evening! gents.

Can someone tell me about the preferable DIMM population for Intel Xeon E5-26xxV2 (Ivy Bridge) if there were these quantities available only:

4 DIMMs 8GB DDR3-10600 ECC (quad channel) per CPU
versus
2 DIMMs 16GB DDR3-14900 ECC (dual channel) per CPU

No specific workload. Dell Precision T3610/5610/7610.
In short:
Does quad-channel operation of slower 1333 MHz RAM compensate somewhat dual-channel operation with faster 1866 MHz RAM?

Regards, Siggi

 

[Sora]

Unless you didn't configured the setup and other files to compatibility mode for it to work on windows 11 what motherboard you trying to use?

nothing to do with compatibility anythings, the iRSTe 8 and lower drivers will not work properly on windows 11

 

[Edwired]

Well if it isn't working on Windows 11 roll back to Windows 10 problem solved. To be honest I don't see the benefits of using Windows 11 on older platform like 1366 unless it was modified to run windows 11 without the possibilities of updates or anything else

 

[Sora]

Well if it isn't working on Windows 11 roll back to Windows 10 problem solved. To be honest I don't see the benefits of using Windows 11 on older platform like 1366 unless it was modified to run windows 11 without the possibilities of updates or anything else

windows 11 is faster than 10, and 10 has an unresolved DXGK composited object leak
the msahci drivers are just as good as the irst's anyway.

 

[Edwired]

Show some reports of that and have a debate about it

 

[Mr.Scott]

windows 11 is faster than 10,

LOL.
Not from a benchmarking point of view. Much slower.

 

[lexluthermiester]

Good evening! gents.

Can someone tell me about the preferable DIMM population for Intel Xeon E5-26xxV2 (Ivy Bridge) if there were these quantities available only:

4 DIMMs 8GB DDR3-10600 ECC (quad channel) per CPU
versus
2 DIMMs 16GB DDR3-14900 ECC (dual channel) per CPU

No specific workload. Dell Precision T3610/5610/7610.
In short:
Does quad-channel operation of slower 1333 MHz RAM compensate somewhat dual-channel operation with faster 1866 MHz RAM?

Regards, Siggi

Go for the quad set. In the Dell systems you listed, all DDR3 runs at 1333mhz. The quad set will allow you to run your system at it's maximum RAM bandwidth. In dual channel mode, you get half RAM bandwidth.

LOL.
Not from a benchmarking point of view. Much slower.

Not in my experience. I have found that it depends on the program/app/game, and generally the differences are VERY minimal.

 

[Sora]

Not from a benchmarking point of view. Much slower.

Now come to reality where you're wrong, my 4Ghz 1680v2 is faster on 11 than 10, in games and benchmarks.

and it doesn't crash every 3-4 days from turning the tv on and off causing explorer to leak handles like a mofo.

If you're doing something stupid like leaving HVCI enabled, than sure its going to be slower.

Show some reports of that and have a debate about it

No can do, Feedback hub doesn't show bugs logged for windows 10 when you're on 11.

Its 100% guaranteed some of you don't know half the things you're talking about when talking performance between 10 and 11, especially those claiming "much slower"

 

[Zyll Goliat]

LOL.
Not from a benchmarking point of view. Much slower.

Honestly they are more or less identical......problem is there are more "bloaters"that are working in the background that need to be stopped but saying that W10 are not much better either it's also full of telemetry BS.....So in short if the system is properly configurated it's going to work faster W10 or W11 doesn't matter.....

 

[maur0]

Now come to reality where you're wrong, my 4Ghz 1680v2 is faster on 11 than 10, in games and benchmarks.

View attachment 289795
View attachment 289796

and it doesn't crash every 3-4 days from turning the tv on and off causing explorer to leak handles like a mofo.

If you're doing something stupid like leaving HVCI enabled, than sure its going to be slower.



No can do, Feedback hub doesn't show bugs logged for windows 10 when you're on 11.

Its 100% guaranteed some of you don't know half the things you're talking about when talking performance between 10 and 11, especially those claiming "much slower"

how do you think windows server 2022 would look on the pc would it win win11?

if you want iso already converted for you to test, the only detail is pt-br but it is possible after installing install language pack for the language you want, I recommend first thing after installing winupdate wheel to update system

Spoiler: iso server 2022 converted workstation pt-br

https://www.mediafire.com/file/vc17...r_convertido_em_desktop_outubro_2022.iso/file

 

[Edwired]

Sorry Sora you can boasts all you want but the benchtest results makes no different in the real world. Your 7 points differences on stock setup don't show the whole story it pretty much the same for any operating system

Its 100% guaranteed some of you don't know half the things you're talking about when talking performance between 10 and 11, especially those claiming "much slower"

That to me tells a lot about you only caring about little gains here or there. We being around a lot longer on here that we know what we are talking about just because you overclocked the cpu and nothing else matters. What you hoping to prove by rubbing a Wishful Thinking Bottle. But the funny thing to notice here is that you posted stock cpu speed results and not the 4ghz cpu speed results

 

[Mr.Scott]

That to me tells a lot about you only caring about little gains here or there. We being around a lot longer on here that we know what we are talking about just because you overclocked the cpu and nothing else matters. What you hoping to prove by rubbing a Wishful Thinking Bottle. But the funny thing to notice here is that you posted stock cpu speed results and not the 4ghz cpu speed results

He's just being a d1ck now. I have no time or patience for that.
I've been an overclocker/competitive bencher for 2 decades. I don't know sh1t.

 

[Edwired]

Same here I have more important things to be doing than bragging about it. About 2 to 3 months nearly moving back home to the house that is nearly finished since the house fire in October 2021 so happy days

 

[SiggiTheHun]

Go for the quad set. In the Dell systems you listed, all DDR3 runs at 1333mhz. The quad set will allow you to run your system at it's maximum RAM bandwidth. In dual channel mode, you get half RAM bandwidth.

Thx for the reply.
However, both "GPU-Z" and "Speccy" show the expectable and correct (half) RAM speed for the installed DDR3-14900 moduls.
Not shure, but think I dreamt or read somewhere that the data rate is capped off when inside Dell T3610 or T7610 all eight DIMMM (2 per channel) slots per CPU are populated.
Can also be -IF!- that this applies to the former T3600/T7600 workstations also/only...

Is "Passmark PerformanceTest" a suitable tool to monitor the product(?) of RAM speed x bandwidth?

Regards, Siggi

 

[Magic Cube]

Couple days ago I did change fan from my Fujitsu's powersupply(same unit as picture). I really haven't been listening it's sound, as I use noise canceling headphones everyday. Reason for fan change was it's loudness. 80mm Delta ball bearing fan with 4300 RPM max speed. Motherboard controls this fan with PWM. Who ever decided this fan's speed profile, should listen this PSU and MB combination until they are dead. It's awful: 1 second full speed, 1 second 1200 RPM, 1 second 900 RPM and back to full speed.

I replaced original fan with Noctua NF-A8 PWM with Low noise adapter->Was only available one with rubber corners and PWM low noise adapter. Had modify Delta's fan connector and low noise adapter, because Delta had non-standard connector. Atleast it is not as loud as before. Max RPM I have seen is about 1400 RPM and similar acting fan speed profile.

Don't have pictures. I wasn't thinking about pictures when i was working on it.

 

[Sora]

how do you think windows server 2022 would look on the pc would it win win11?

if you want iso already converted for you to test, the only detail is pt-br but it is possible after installing install language pack for the language you want, I recommend first thing after installing winupdate wheel to update system

Spoiler: iso server 2022 converted workstation pt-br

https://www.mediafire.com/file/vc17...r_convertido_em_desktop_outubro_2022.iso/file

This is actually based on one of the final windows 10 builds, prior to the switch to windows 11 branding, it has the features of 11, but the 10 start menu and bar.

I've been an overclocker/competitive bencher for 2 decades. I don't know sh1t.

People citing terms of doing anything tend to be bignoters that know nothing and travel with the flock.
Sorry, but you're claims of reduced performance indicates an issue with your machine and testing under disimilar environment configurations, not an issue with 11 vs 10 itself

Sorry Sora you can boasts all you want but the benchtest results makes no different in the real world. Your 7 points differences on stock setup don't show the whole story it pretty much the same for any operating system

I don't think you even realise what a 7 point single thread increase even constitutes, more importantly is the big step up with the multithread score where 100ish points would typically correlate to an extra 200mhz on the core clocks. My 1680v2 is performing as if its running 4.6ghz rather than 4.2 by just upgrading from 10 to 11, though that is a take based on the improvements from the Bulldozer scheduler patch against Nahalem and R11 from back when that happened.

 

[maur0]

This is actually based on one of the final windows 10 builds, prior to the switch to windows 11 branding, it has the features of 11, but the 10 start menu and bar.



People citing terms of doing anything tend to be bignoters that know nothing and travel with the flock.
Sorry, but you're claims of reduced performance indicates an issue with your machine and testing under disimilar environment configurations, not an issue with 11 vs 10 itself




I don't think you even realise what a 7 point single thread increase even constitutes, more importantly is the big step up with the multithread score where 100ish points would typically correlate to an extra 200mhz on the core clocks. My 1680v2 is performing as if its running 4.6ghz rather than 4.2 by just upgrading from 10 to 11, though that is a take based on the improvements from the Bulldozer scheduler patch against Nahalem and R11 from back when that happened.

one thing that I do and it helps a lot on the cpu of our ivy bridge line is to turn off the specter and meltdown protections, it can lose up to 20% of performance if activated, do you leave it deactivated there too? if it is activated, deactivate it and run benchmark to see if it gains some performance
to disable I leave the registry keys

Windows Registry Editor Version 5.00

[HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Control\Session Manager\Memory Management]
"FeatureSettingsOverride"=dword:00000003
"FeatureSettingsOverrideMask"=dword:00000003

https://www.grc.com/files/InSpectre.exe

 

[Edwired]

one thing that I do and it helps a lot on the cpu of our ivy bridge line is to turn off the specter and meltdown protections, it can lose up to 20% of performance if activated, do you leave it deactivated there too? if it is activated, deactivate it and run benchmark to see if it gains some performance
to disable I leave the registry keys

Windows Registry Editor Version 5.00

[HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Control\Session Manager\Memory Management]
"FeatureSettingsOverride"=dword:00000003
"FeatureSettingsOverrideMask"=dword:00000003

https://www.grc.com/files/InSpectre.exe

I wouldn't be fiddling with the registry as you be leaving your computer exposed to those exploits. I would just overclock the system to gain back some percentage of the loss in performance and still be protected to those exploits

 

[lexluthermiester]

Thx for the reply.

You're welcome! BTW, welcome to TPU!

Not shure, but think I dreamt or read somewhere that the data rate is capped off when inside Dell T3610 or T7610 all eight DIMMM (2 per channel) slots per CPU are populated.

I've never read or seen that. They should run quad-channel, unless you install mis-matched DIMMs. If they're all matched, they'll run quad-channel.

one thing that I do and it helps a lot on the cpu of our ivy bridge line is to turn off the specter and meltdown protections, it can lose up to 20% of performance if activated, do you leave it deactivated there too? if it is activated, deactivate it and run benchmark to see if it gains some performance
to disable I leave the registry keys

Windows Registry Editor Version 5.00

[HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Control\Session Manager\Memory Management]
"FeatureSettingsOverride"=dword:00000003
"FeatureSettingsOverrideMask"=dword:00000003

https://www.grc.com/files/InSpectre.exe

I always disable those patches. They are useless. There are no known(or unknown) exploits in the wild. Put another way, while these vulnerabilities exist, they so immensely difficult to remotely exploit that it is effectively impossible, even with physical access it's still VERY difficult.

 

[Sora]

one thing that I do and it helps a lot on the cpu of our ivy bridge line is to turn off the specter and meltdown protections, it can lose up to 20% of performance if activated, do you leave it deactivated there too? if it is activated, deactivate it and run benchmark to see if it gains some performance

Retpoline reclaimed 98% of performance on Windows 10+, Which I do have enabled on the machine, Import Optimizations is presently off however as my virtual disc software hasn't found a way to work without doing so prsently.

 

[Edwired]

Xeon Black Ops cpu as fully 6 core no threads

 

[lexluthermiester]

Xeon Black Ops cpu as fully 6 core no threads

Crazy voltage! I would have dropped it right off. 1.25 to start off. 4.6ghz really shouldn't need 1.55v.