Xeon Owners Club

[Edwired]

may it rest in peace.

It a crispy death for a chip

 

[maur0]

I was having a series of bugs in the USB asmedia 3.0 where the PC hangs after leaving standby, the joystick pressing the button alone etc etc and after updating the USB firmware the 64kb file solved all these annoying problems

update for asmedia usb 3.0 ASM1042 xHCI controler, instaled on asus rampage IV extreme
When updating here I realized that I was using a USB flash drive from 2011, this one is from 2013, this update is outside the BIOS, even though the BIOS is the latest from 2014, the USB firmware has not been updated since it left the factory in 2011, full of bugs.

After updating and restarting the PC, wait about 2 minutes as USB devices will work. It took this long to detect the mouse after updating firmware

download

Asus USB 3.0 FW Update Tool v3.1

MediaFire is a simple to use free service that lets you put all your photos, documents, music, and video in a single place so you can access them anywhere and share them everywhere.

www.mediafire.com

I found this site, there are others which one you think is best to use on the site, it seems that the last version is this one for the HP ASM104x v1.5.2.7 FW131025-10-11-AB, will it work correctly on the Asus Rampage IV?

https://www.usbdev.ru/files/asmedia/104xfwupg/

I'm afraid of installing this and my USB 3.0 won't work anymore hehe is there any chance of this happening using this HP firmware or not?

Who updates to this from HP on the Asus Rampage IV Extreme, please report if it was good


Did you update the USB controller firmware on your motherboard?

 

[xtreemchaos]

"he who dare's wins" but what do i know eh. there's allways restore points even with bios just restore og settings

 

[Edwired]

The symptom appears as the USB 3.0 is flaky giving all the problems surely firmware update may fix it or brick it. What you could have done is turn it off and get a pci-e usb card problem solved

I was having a series of bugs in the USB asmedia 3.0 where the PC hangs after leaving standby, the joystick pressing the button alone etc etc and after updating the USB firmware the 64kb file solved all these annoying problems

update for asmedia usb 3.0 ASM1042 xHCI controler, instaled on asus rampage IV extreme
When updating here I realized that I was using a USB flash drive from 2011, this one is from 2013, this update is outside the BIOS, even though the BIOS is the latest from 2014, the USB firmware has not been updated since it left the factory in 2011, full of bugs.

After updating and restarting the PC, wait about 2 minutes as USB devices will work. It took this long to detect the mouse after updating firmware

download

Asus USB 3.0 FW Update Tool v3.1

MediaFire is a simple to use free service that lets you put all your photos, documents, music, and video in a single place so you can access them anywhere and share them everywhere.

www.mediafire.com

I found this site, there are others which one you think is best to use on the site, it seems that the last version is this one for the HP ASM104x v1.5.2.7 FW131025-10-11-AB, will it work correctly on the Asus Rampage IV?

https://www.usbdev.ru/files/asmedia/104xfwupg/

I'm afraid of installing this and my USB 3.0 won't work anymore hehe is there any chance of this happening using this HP firmware or not?

Who updates to this from HP on the Asus Rampage IV Extreme, please report if it was good


Did you update the USB controller firmware on your motherboard?

 

[NeoThuNder]

Hai Guys,
I had old motherboard with socket LGA1155. if i remember right, its a H61 and currently using Intel G2020 ( the processor already dead ). i want to replace it to Xeon.
Which one is better ?
cause its my first time to pick Xeon as my build.

 

[uco73]

i want to replace it to Xeon.
Which one is better ?

For 1155 socket are Xeons E3 series.

https://www.cpu-world.com/Sockets/Socket%201155%20(LGA1155).html

 

[maur0]

im testing server 2025 RTM on old ivy bridge xeon, work Cool, even though it doesn't have "official" support from MS, it runs perfectly, great performance for the 2013 cpu

 

[uco73]

im testing server 2025 RTM on old ivy bridge xeon, work Cool, even though it doesn't have "official" support from MS, it runs perfectly, great performance for the 2013 cpu

E5-2697v2 is still the beast!

 

[sk111]

Hi has anyone used ThrottleStop to fine tune undervolt settings after Turbo Boost Unlock of a xeon 2600 v3 series processor?
I have successfully done TurboBoost unlock process on a 2697 v3 so the processor is fully unlocked at following undervolts -50/-50/-30 (last for system agent) and it runs at 3.6GHz and I would like to modify these 3 undervolts for stability testing. However, as soon as I tick unlock adjustable voltages in ThrottleStop the processor speed drops immediately to 3.1 GHz from 3.6 GHz even if I do not touch the voltages at all (as measured in hwmonitor and CPUZ). Any ideas what is happening and if I can use Throttlestop or is it a bug of the software? See screenshots below.
Pic 1: Stress test before TS
Pics 2-4: Settings correctly picked up in TS
Pic 5: Ticking apply adjustable voltage and hitting apply without changing anything
Pic 6: Stress test after above. Clock drops to 3.1GHz and performance reduction

 

[:D:D]

@sk111 see https://forums.anandtech.com/threads/what-controls-turbo-core-in-xeons.2496647/post-38697001

BTW be careful of other software undoing the change, you may need to lock MSR 0x194 (see previous posts). For instance Throttlestop used to reset this type of OC if microcode update was in place, I don't know if that has changed since.

IOW to do that you would need to not have any microcode update present IINM, note that windows usually updates microcode too. Having no microcode might bring it's own problems. Usually a positive offset is used for SA, it doesn't affect CPU ratio's but may possibly help with memory and bclk AFAIK.

 

[sk111]

@sk111 see https://forums.anandtech.com/threads/what-controls-turbo-core-in-xeons.2496647/post-38697001


IOW to do that you would need to not have any microcode update present IINM, note that windows usually updates microcode too. Having no microcode might bring it's own problems. Usually a positive offset is used for SA, it doesn't affect CPU ratio's but may possibly help with memory and bclk AFAIK.

Hey thanks. Unclewebb confirmed in another thread that TS cannot be used on this occasion as it seems to be resetting the core speed so will leave that for now. Will have to do testing by redoing the TU process a couple of times to see what is optimal in terms of speed/stability

 

[:D:D]

What board? Do you not have BIOS options to adjust voltage? Personally I undervolt my 2696 v3 core and cache ust to the point where an extra bin can be had. Your 3.5GHz all core sounds good for all 14 cores. Cannot remember what my 18 core gets when set to 14 cores but IIRC it wasn't until enabling only 10 cores I got 3.8GHz all core and I think ~3.7GHz with AVX2/FMA.

 

[Edwired]

 

[xtreemchaos]

great vid bud, there's life in the old dogs. the MR9A pro board he showed is one the best China boards i've come across so far with 8 phases it runs my 2699 well no probs with it yet "fingers crossed" every mainstream board bar one has kicked the bucket on me but i've only had 3 to be truthful and they was all 2hand and prob well hammered before i got them. thanks.

 

[lexluthermiester]

Just a video? No commentary? Come on, do share. What are your thoughts on the perspective offered in that video?

 

[xtreemchaos]

my take is good overclock but a lot of power but again its good fun but for every day use id stick to stock i mean 3.8 isnt bad with 8 cores, i get 3.6 with my 18c with the boost mod i havn't a clue how much power it uses but the misses hasn't complained about using too much power .

 

[budget_Optiplex]

Love me some X99 so thanks for the video that I'll be watching shortly! I built my system for literally next to nothing and am quite happy with it's performance. Though I just switched from a Broadwell-E 1650 v4 to a Haswell-E 2643 v3 as it has 5MB more L3 cache and actually maintains more consistent and higher all core turbo clocks - and does perform better in the benchmarks I have ran probably because of the additional cache. I also am seriously unimpressed with Turbo Boost 3.0 available in some Broadwell-E 16xx CPUs and the lack of driver support due to supposed security problems, thusly I choose the Haswell-E with good ol Turbo Boost 2.0 instead. I was going to try a Broadwell-E 2643 v4 but those are still $25 versus the $7 I paid for the Haswell v3 lol! I like my computer parts to cost less then a trip to a restaurant lol!

Someday I may solder a socket on this Lenovo board so I can mod the BIOS with the Haswell Turbo hack and be able to burn the chips with my burner and easily swap them in until I figure out how to work around the Lenovo BIOS. I have not seen anyone yet do the Turbo hack to a Lenovo board.

 

[lexluthermiester]

my take is good overclock but a lot of power but again its good fun but for every day use id stick to stock i mean 3.8 isnt bad with 8 cores, i get 3.6 with my 18c with the boost mod i havn't a clue how much power it uses but the misses hasn't complained about using too much power .

I haven't bothered with much OCing in a few years. I do a bit here and there, but it's just not worth it like it once was.

As for the old Xeon V3 & V4 models, they are still decent value but the points made in that video above are not to far out of line. Still, I would rather have an 8core V3 Xeon than a 12100f any day of the week for many reasons.

 

[xtreemchaos]

8core V3 Xeon than a 12100f any day of the week for many reasons.

me too. nower days it plug n play unless on LN2 or some other. I have HT turned off just to save a bit of power and help maintain clocks

Someday I may solder a socket on this Lenovo board so I can mod the BIOS with the Haswell Turbo hack

Where's there's a will there's a way. i was chuffed when i got rebar working on mine with using a a770.

 

[lexluthermiester]

I have HT turned off just to save a bit of power and help maintain clocks

I do the same with HT, but more to mitigate the Spectre/Meltdown/Side-channel type vulnerabilities, which most(not all) depend on HT being enabled. The moment a user turns off HT, most of those vulnerabilities disappear. The performance difference isn't much to be concerned with, especially given that the boost clocks staying higher mostly offset the loss of threads.

 

[xtreemchaos]

Spectre/Meltdown/Side-channel type vulnerabilities,

i didn't know that bud, that make's HT off even more of a +. yes the performance loss is small if i remember right in CB its about 20% with about 35% power saving " my guess" because its hard to pin down with boards from china.

 

[Edwired]

I didn't expect an outburst of activity like this to happen was only posting an interesting video based on xeon set up. I have not yet upgraded to the x79 or x99 yet because I haven't my system with me. Still on the lookout for a place of my own so it be awhile

 

[:D:D]

@sk111 tried running with just 14 cores enabled.

So hopefully you can get full 36x ratio on all cores with non FMA/AVX2 for which I get 34x and didn't try to get more.

 

[budget_Optiplex]

I just got around to a project I intended to do when I first got my Haswell-E E5-2643 v3. I have noticed many Haswell-E heat-spreaders have a noticeable 'bump' around the vent hole. My example was particularly bad (see example picture - not my actual CPU), so I plugged the vent hole with thermal paste, broke out some 600 and 1200 grit wet type sandpaper and got to work doing some lapping then lots of cleaning of the CPU. It was a quick job as anything was an improvement over what I started with but it made a big difference in temps and I suggest if running a Haswell-E this issue may need addressed. For some reason I haven't seen this 'bump' on Broadwell-E CPUs.

 

[Dr. Dro]

Hello friends, I want your input before I make a decision. As you know, Windows is kind of going to hell in a handbasket lately and with its future being uncertain, I'm feeling a little unsafe regarding having only a latest generation computer running the latest version of Windows 10/11 available to me. So, with the objective of keeping a fully functional Windows 7 system around, I decided to dig out the parts from my drawer and start a "new" build. This is to be used as a general purpose desktop; media consumption, light gaming, some casual video encoding. It's basically what I do with my PC but with lower expectations.

I own a Gigabyte X99 Ultra Gaming motherboard currently equipped with a 18-core Xeon E5-4669 v3 CPU. My complaint is that the clock speeds on this processor really suffer. It wasn't as bad when I first acquired it back in 2017 or so, but nowadays apps require a level of single thread well beyond what it can do - even after pulling the turbo unlock trick it doesn't do more than 2.9 GHz (or 2.4 GHz if AVX is enabled, this is a hardcoded offset), it's woefully slow. It will be paired with the RAM kit that used to be part of my old 5950X system and a GTX 1070 Ti GPU that I keep as a backup/spare card.

So, I've decided to search for a new processor. I've set a small budget towards it and would prefer it to feature an upgrade to Broadwell, I did some research and the options I have are essentially as follows:

E5-1650 v4 (6-core. identical to Core i7-6850K but locked, 3.6 base, ~4.0 turbo) ~$30
E5-2697A v4 (16-core, 2.6 base, ~3.6 turbo) ~$33

-or Haswell options-

E5-2699 v3 (it is identical to the processor I currently have, but up to 700 MHz faster, so up to 3.6 or so - likely subject to the same AVX offset issue, though) ~$57
Core i7-5960X (8-core unlocked, but also Haswell, and easily the one that will achieve the highest clocks) ~$62

My question is, with this objective in mind, which one of these would you personally pick? I'm currently leaning towards the 2697A v4, it seems the most balanced option here, it is cheap, I get the Broadwell upgrade (slightly higher IPC, TSX support), the extra clock frequency I want while losing two cores, and it looks like ten extra cores might as well outweigh losing 400 MHz over the 1650 v4, even with the topology concerns.

One last request if a mod sees this, can you check if I'm "shadowbanned" from this thread (if such a thing exists)? I seem to recall that a couple of years ago when I was still a greenhorn on the forum I had a little of an argument here, and one or two of my posts were flagged as low-quality. I haven't received notifications about this thread ever since.