X299 Owners Club

[tps3443]

yeah u gotta give it lil try cuz its way deeper then just Multiplier OC the normal way without touching the bclock, sometines u just only need to put the bclk higher and thats it, no need to oc further hehe ^^


yahh well degredation happends but much slower then we really like to admit, sure, especially thoes crapzy chips but they will prolly degrade at a time where that machine that 7980 or what ever is useless to you as a main System prolly move it down as second pc and downclock it and get rid some of the voltage and be happy, but degradation hapends sadly everywhere, with CPUs tho idk you will use one for that long where its going to be a problem,,, prolly is the Performance which is before its a degredation but we can all agree at that level of hardware, almost every degree is a win

have a good morningy dudes, ive gott clean my sidepanel soon hehe ^^

thehe ^_^

i might still go for the 10980xe maybe but my 10940x already decasifies the 7980xe but it was too good for its time still so if somebud has it, optimise it a lil and put an alphacool eisbaer or some in that range on it and yeah cary on and have fun

I owned a 10980XE for a brief period, other than better silicon. My 10980XE and 7980XE performed the same.

I ran the 7980XE on an older bios with older ME and direct die before specter melt down and all that, so essentially it was actually faster than my 10980XE because the 7980XE wasn’t just out clocking it, but it was faster per clock without the newer mitigations in place on the 10980XE. I left the 10980XE soldered, and eventually returned it for a refund, mainly because I could not afford to keep both.

You may be talking about stock performance of these chips though. They did boost up the stock 10980XE quite a bit over the 7980XE. But I always looked at the 7980XE/9980XE/10980XE as the same chips for some reason other than solder.

The 7980XE had a weak stock “All-Core boost” I believe it did 3.4Ghz with (18/36) loaded.

 

[Zyll Goliat]

I owned a 10980XE for a brief period, other than better silicon. My 10980XE and 7980XE performed the same.

I ran the 7980XE on an older bios with older ME and direct die before specter melt down and all that, so essentially it was actually faster than my 10980XE because the 7980XE could out clock it. I left the 10980XE soldered, and eventually returned it for a refund.

I’m running an older ME/bios on my 3175X as well. You may be talking about stock performance of these chips

Well I tried to explain that to him few pages back but seems like he don't get it 7,9,10 series on X299 are basically the same except for the stock/turbo clocks....
Anyway I have 7940X on Asus Rog Gaming and still using older bios because on that bios I can use ECC ram and currently have 192Gb that works on 2400Mhz and I managed to tweak a bit timings so aint that bad in quad channel..... @tps3443 that Xeon 3175X is true MONSTER how long you have it and is it still expensive and what about the mobo that's LGA 3647 if I am not wrong are those still very pricey?

 

[tps3443]

Well I tried to explain that to him few pages back but seems like he don't get it 7,9,10 series on X299 are basically the same except for the stock/turbo clocks....
Anyway I have 7940X on Asus Rog Gaming and still using older bios because on that bios I can use ECC ram and currently have 192Gb that works on 2400Mhz and I managed to tweak a bit timings so aint that bad in quad channel..... @tps3443 that Xeon 3175X is true MONSTER how long you have it and is it still expensive and what about the mobo that's LGA 3647 if I am not wrong are those still very pricey?

Yeah the 7980XE was running at a very tame 3.4Ghz on all 18 cores stock. It was efficiently tuned though. The 10980XE ran +400Mhz faster, and even this Xeon 3175X stock during an 18/36 core load runs 4.0Ghz. So lots of chips out of the box will or can potentially whoop the stock 7980XE. But once it’s overclocked, it changes drastically. My 7980XE wasn’t a fantastic sample by any means. But they all have HUGE OC potential.

But yeah the 3175X and LGA 3647 are still kinda expensive to build nowadays. I’ve only had mine for 4+ months now. I actually had a 14900KS/Z790 Apex before this. And I went out and bought a Dominus Extreme, and a Xeon 3175X to play with and use daily, and I traded off the 14th gen Intel system. This system has been a very reliable platform though. As for ram, I know some of those 7th gen HEDT chips can run some configs that aren’t even supported or supposed to even work. I also heard the ram limits on these chips are much higher than what’s actually listed as well. Intel was literally just saying a ram max on the specs, and not actually locking things down. I have seen someone running 1TB or even more ram with a 3175X but apparently and they are 512GB/768GB ram limits which isn’t true at all. ‍

 

[Zyll Goliat]

Yeah the 7980XE was running at a very tame 3.4Ghz on all 18 cores stock. It was efficiently tuned though. The 10980XE ran +400Mhz faster, and even this Xeon 3175X stock during an 18/36 core load runs 4.0Ghz. So lots of chips out of the box will or can potentially whoop the stock 7980XE. But once it’s overclocked, it changes drastically. My 7980XE wasn’t a fantastic sample by any means. But they all have HUGE OC potential.

But yeah the 3175X and LGA 3647 are still kinda expensive to build nowadays. I’ve only had mine for 4+ months now. I actually had a 14900KS/Z790 Apex before this. And I went out and bought a Dominus Extreme, and a Xeon 3175X to play with and use daily, and I traded off the 14th gen Intel system. This system has been a very reliable platform though. As for ram, I know some of those 7th gen HEDT chips can run some configs that aren’t even supported or supposed to even work. I also heard the ram limits on these chips are much higher than what’s actually listed as well. Intel was literally just saying a ram max on the specs, and not actually locking things down. I have seen someone running 1TB or even more ram with a 3175X but apparently and they are 512GB/768GB ram limits which isn’t true at all. ‍

Well.....that "monster"of yours supports a hexa/six channel memory and you have there 12 slots so if that mobo supports ecc ram and if you could find those rare and expensive 128Gb ddr4 sticks then I really can't see the reason why should not work and theoretically then it could support 12x128Gb=1536Gb(1,5Tb)of ram......well same goes for X299 it's just that have few slots(8)less and it's quad channel instead but realistically it's much easier and much cheaper to get 32Gb sticks or even maybe 64Gb so my guess that you can get 768Gb that's not going to cost you anymore yours "arms&legs".......
BTW what are you doing on that machine is it a "working-horse" or just a enthusiast fun project?

P.S.I am still OK with this 7940X 'tho I was thinking about 7980XE as possible upgrade but I am really not in a hurry as this machine is more than enough for my needs atm....

 

[Corralz]

So i upgraded my 7900x to my 9920x i bought and i so far its been a better than expected. Ive been able to get .2 ghz higher overclock so far and about 10 degrees colder while benchmarking and 10-15 degrees colder while gaming. So far at 4.5ghz i can run cinebench r24 but on r23 it crashes. Im still on stock voltage so i may mess with that a little bit. Its at 1.17v
Ill be happy with a 4.6 all core and maybe 4.8 turbo on the 2 coldest cores.

 

[lexluthermiester]

So i upgraded my 7900x to my 9920x i bought and i so far its been a better than expected.

Nice upgrade! While it's only 2 cores, there's also some speed bump and the IPC improved as well from the revision.

Im still on stock voltage so i may mess with that a little bit. Its at 1.17v

You should only need a slight bump, maybe 1.19 or 1.20 to get there.

 

[tps3443]

So i upgraded my 7900x to my 9920x i bought and i so far its been a better than expected. Ive been able to get .2 ghz higher overclock so far and about 10 degrees colder while benchmarking and 10-15 degrees colder while gaming. So far at 4.5ghz i can run cinebench r24 but on r23 it crashes. Im still on stock voltage so i may mess with that a little bit. Its at 1.17v
I’ll be happy with a 4.6 all core and maybe 4.8 turbo on the 2 coldest cores.

Could you do a comparison between these chips. Voltages needed. Power consumption required etc.

 

[Corralz]

Could you do a comparison between these chips. Voltages needed. Power consumption required etc.

I rem

Nice upgrade! While it's only 2 cores, there's also some speed bump and the IPC improved as well from the revision.

You should only need a slight bump, maybe 1.19 or 1.20 to get there.

i was able to get it stable at 4.5 ghz with 1.19v , I tried to get 4.6 ghz but i need 1.25v to get it to run c24. I’m scared to degrade my chip so I may just run it at 1.20 and do a custom oc on the colder cores . And the keep the hottest cores at 4.5

 

[Zyll Goliat]

I rem

i was able to get it stable at 4.5 ghz with 1.19v , I tried to get 4.6 ghz but i need 1.25v to get it to run c24. I’m scared to degrade my chip so I may just run it at 1.20 and do a custom oc on the colder cores . And the keep the hottest cores at 4.5

Maybe you should try to OC your CPU individually by core I managed to OC my 7940X like that and I have 2 cores OC on 4,7Ghz with only 1,218V....it is a tricky and you will need a lot of time to test but it could be rewarding my cores works from 4,1Ghz all the way up to the 4,7Ghz and the results are pretty good for me....

 

[C7TJFVB7-CAZ]

ya can all hope for an External Clock Gen, changes everything by gigving everything but the SATA/ IO 25mhz more
4.125ghz is just as fast as 4.8ghz on 100mhz bclock but you sadly gotta have the right motherboard though :/

 

[tps3443]

ya can all hope for an External Clock Gen, changes everything by gigving everything but the SATA/ IO 25mhz more
4.125ghz is just as fast as 4.8ghz on 100mhz bclock but you sadly gotta have the right motherboard though :/

So, I don’t think this is the case. You are merely overclocking the CPU different way. The end result is the same thing. Just different techniques.

If my CPU is at 4.8Ghz with a 100Mhz BCLK, then it’s gonna be way faster than my CPU would be at 125Mhz BCLK with a x33 multiplier which would equate to 4.125Ghz.

I watched the video you shared on clock generator. And this only came in handy with cpu’s that have a locked multiplier.

If we have an unlocked CPU, then BCLK overclocking is only using for certain things. For example squeezing out the last bit of an overclock. I can run 4.9Ghz but not 5.0Ghz, so I can use a 105Mhz with x49 to give me 4,950Mhz which may very well be the stable limits. Or, if I’m trying to achieve a specific memory speed, or slightly more memory overclock. I can run DDR4 3800 cl14, but I can’t run DDR4 4000. So I may use a 125bclk, or other bclk with a divider so I can achieve something like 3866, or maybe 3933Mhz DDR4. Or even 3900Mhz.

But using a 125 BCLK isn’t really giving me an advantage at all other than running a 125bclk just to achieve a different end result.

But using a 125bclk isn’t going to really offer any performance advantage that a system with a 100bclk can. It’s just giving locked CPU’s a chance to run faster. And or more fine tune in dialing in the extreme limits of an overclock.

 

[C7TJFVB7-CAZ]

well sadly youre a bit wrong, otherwise tell my why my 4080 is utilised more then at 100mhz busspeed hm? effectively 25% more IPC on that thing and 25% more bandwith to the pcie slot, and a card with gen4 on a gen 3 need all the more bandwith you can give it
PS i should have said this is for the 40 series cince they WAAY fastwer and different then the 30 series soooo yeah, if 3440x1440p is hard on a 4080, you know its ya CPU and theres no one who can convince me otherwise, maybe you guys got a 6800XT or idk 30 series where NO! theere it wont make a diffrence in gaming but everythign else tho, maaybe you guys just dont DO as much on ya machines where you wont see if the pc is really faster by using iut wehre you will need a benchmark to tell the difference

 

[tps3443]

well sadly youre a bit wrong, otherwise tell my why my 4080 is utilised more then at 100mhz busspeed hm? effectively 25% more IPC on that thing and 25% more bandwith to the pcie slot, and a card with gen4 on a gen 3 need all the more bandwith you can give it

So, I tried this my self after you said something before. I tested 3DMark PCIe bandwidth test using stock 100Mhz BCLK, and then tried again with a higher BCLK. I did not see any difference in the amount of bandwidth/GB/s that the PCIe slot provides if it is running beyond the 100Mhz.

You are welcome to try your self and share your results.

I’m throwing a 5090 on this thing lol. So I wish it did help.

 

[C7TJFVB7-CAZ]

So, I tried this my self after you said something before. I tested 3DMark PCIe bandwidth test using stock 100Mhz BCLK, and then tried again with a higher BCLK. I did not see any difference in the amount of bandwidth/GB/s that the PCIe slot provides if it is running beyond the 100Mhz.

You are welcome to try your self and share your results.

I’m throwing a 5090 on this thing lol. So I wish it did help.

look at that singlecore baby i dont benmchmark a lot but this ones thankfully closer to reality since ive sat down on BOTH machines they feel the same to me, did i waste my time like you dudes and OC over the multiplier i wont even get that close and neither do i see my GPU over like 65% usage WOOW how great oof you dont seem to get it idk how do you dudes want me to explain this? :/ its all about benchmarking for your dudes that basicaly ALL you guys do thats it, rather then actually have friggin sense about anything
dudes always seem to make me a bit mad on perpose, cuz im a realworld usage kinda guy, waaay less about stupid fricking benchmarks that they often wont even tell the full story, which is why 8 cores still sell today like warm bread aka 5800x3D/7800x3d cuz they saw a benchmark chart and think this would always apply to them even tho they where tested with a 3090/4090 at 1080p/ slap in anything less then a 90 card, or dotn have 1080 and most of that gap goes away where u could just as well have bought yourself an R7 7700X because it would be preforming pretty much exactly the same :*
and i see this everytime, ive had one dude on a 3090 at 1440p and 2 moinitors had a 5800x3d paired together, took a long time for me to explain him that a more core cpu is better and indeed he bought the right cpu then which was the 5950x and he THEN was gettn all the power of a 3090, couldnt beleave it becasue he was lead by benchmarks rather then, well you know user experience

skip to 12:50 so u know that im talking about he might be better at explaining then me

 

[tps3443]

look at that singlecore baby i dont benmchmark a lot but this ones thankfully closer to reality since ive sat down on BOTH machines they feel the same to me, did i waste my time like you dudes and OC over the multiplier i wont even get that close and neither do i see my GPU over like 65% usage WOOW how great oof you dont seem to get it idk how do you dudes want me to explain this? :/ its all about benchmarking for your dudes that basicaly ALL you guys do thats it, rather then actually have friggin sense about anything
dudes always seem to make me a bit mad on perpose, cuz im a realworld usage kinda guy, waaay less about stupid fricking benchmarks that they often wont even tell the full story, which is why 8 cores still sell today like warm bread aka 5800x3D/7800x3d cuz they saw a benchmark chart and think this would always apply to them even tho they where tested with a 3090/4090 at 1080p/ slap in anything less then a 90 card, or dotn have 1080 and most of that gap goes away where u could just as well have bought yourself an R7 7700X because it would be preforming pretty much exactly the same :*
and i see this everytime, ive had one dude on a 3090 at 1440p and 2 moinitors had a 5800x3d paired together, took a long time for me to explain him that a more core cpu is better and indeed he bought the right cpu then which was the 5950x and he THEN was gettn all the power of a 3090, couldnt beleave it becasue he was lead by benchmarks rather then, well you know user experience

skip to 12:50 so u know that im talking about he might be better at explaining then me

How about this. Go run Cinebench of your choice Single thread with x100, then go run with x125. CPU speed should be the same between runs.

 

[C7TJFVB7-CAZ]

How about this. Go run Cinebench of your choice Single thread with x100, then go run with x125. CPU speed should be the same between runs.

pretty much would be yeah but not today bro ill do it when ever i dont need to use my machine but im mostly used to R15 whcih isnt a good test anymore for this system xcuz i dont know how it releates to older CPUs we are muc hmore used to. 3th to idk 9th gen intel and 1th to 6 gen ryzen im a bit rusty there
sadly you didnt relayed to the other sstuff i said so you only want to get YOUR point proven, so i might lookout for annother forum, you just benchmark too much for me ill post some crap annother time

 

[Zyll Goliat]

well sadly youre a bit wrong, otherwise tell my why my 4080 is utilised more then at 100mhz busspeed hm? effectively 25% more IPC on that thing and 25% more bandwith to the pcie slot, and a card with gen4 on a gen 3 need all the more bandwith you can give it
PS i should have said this is for the 40 series cince they WAAY fastwer and different then the 30 series soooo yeah, if 3440x1440p is hard on a 4080, you know its ya CPU and theres no one who can convince me otherwise, maybe you guys got a 6800XT or idk 30 series where NO! theere it wont make a diffrence in gaming but everythign else tho, maaybe you guys just dont DO as much on ya machines where you wont see if the pc is really faster by using iut wehre you will need a benchmark to tell the difference

It's because when you do BCLK/OC then you are not OC only the CPU you also OC the PCI e slot and memory....it will work faster for sure but usually if you rise the BCLK a lot ain't stable at all and sometimes could be potentially dangerous for hardware depends of the platform I did BCLK OC on X79 successfully and without a problem with 115 BCLK and it was stable 100% like that for years but on X299 even after like 103 on bus it's not stable in tests same things goes for X99 so depends....for just few % it's not worth the risk in my opinion especially on X299 where you can easily OC any core individually and play with the core voltages separately.....

P.S.You keep posting CPU-Z test over and over again because that is the only test that works on that high BCLK frequency because it's light and stupid simple test go and do Cinebench 23 or 24 and post your results or any other a bit heavier benchmark....well I bet you can't because your Rig ain't stable and you already confirm that in some previous post.......

 

[tps3443]

It's because when you do BCLK/OC then you are not OC only the CPU you also OC the PCI e slot and memory....it will work faster for sure but usually if you rise the BCLK a lot ain't stable at all and sometimes could be potentially dangerous for hardware depends of the platform I did BCLK OC on X79 successfully and without a problem with 115 BCLK and it was stable 100% like that for years but on X299 even after like 103 on bus it's not stable in tests same things goes for X99 so depends....for just few % it's not worth the risk in my opinion especially on X299 where you can easily OC any core individually and play with the core voltages separately.....

P.S.You keep posting CPU-Z test over and over again because that is the only test that works on that high BCLK frequency because it's light and stupid simple test go and do Cinebench 23 or 24 and post your results or any other a bit heavier benchmark....well I bet you can't because your Rig ain't stable and you already confirm that in some previous post.......

I mean, the main reason why I say for him to use another single thread test is because I can easily manipulate CPUz scores. They aren’t exactly consistent, or very reliable.

But single thread Cinebench should be easily to run. Only 1 core is being used. He can probably run that even if it’s extremely unstable lol. I just want to see before and after 100bclk vs 125bclk. And see if his single thread or IPC is actually
Picking up even if the cpu is running at the same speed.

That would be pretty crazy to see.

 

[Corralz]

I mean, the main reason why I say for him to use another single thread test is because I can easily manipulate CPUz scores. They aren’t exactly consistent, or very reliable.

But single thread Cinebench should be easily to run. Only 1 core is being used. He can probably run that even if it’s extremely unstable lol. I just want to see before and after 100bclk vs 125bclk. And see if his single thread or IPC is actually
Picking up even if the cpu is running at the same speed.

That would be pretty crazy to see.

Yeah that’s crazy because my 9920x is putting nowhere near those cpu z numbers . I got my single core at 4.7 ghz right now

 

[tps3443]

Yeah that’s crazy because my 9920x is putting nowhere near those cpu z numbers . I got my single core at 4.7 ghz right now

I didn’t see any difference my Dominus Extreme does pretty well with BCLK overclocking. I just don’t really need to do it, I use the multiplier. I might use a 100.20 BCLK daily for even numbers under load and consistency. But that’s about it.

 

[lexluthermiester]

I might use a 100.20 BCLK daily for even numbers under load and consistency. But that’s about it.

You mean 102.00 bclk?

 

[Gigaherz]

No, 100.2 is to avoid the 99s and get even readouts

 

[lexluthermiester]

No, 100.2 is to avoid the 99s and get even readouts

Wasn't aware any motherboard had that level of fine-grained controls.

 

[Gigaherz]

Its a thing since Uefis exist. So like 2011?

 

[tps3443]

Something interested in games has been happening lately, so I wanted to share. Typically I expect a slight performance loss from my prior 14900KS.

I used to run a 14900KS with DDR5 8600c36 daily, and this was a ripper of a CPU it was tuned to the max and pushed my 4090 extremely well. I’m a fanatic over 1% lows, and GPU usage etc. Anyways I’d be lying if I said all of my games performed better on my Xeon w3175x (In fact several of them perform worse)

However, I have come across at least (3) games where my experience and performance is better than my 14900KS ever was.

(1) Microsoft Flight Simulator
(2) Indiana Jones and the Great Circle
(3) Starwars Outlaws

I run my Xeon 3175X at 4.9Ghz/3.3Ghz mesh/DDR4 3800cl14/ HT=On. And this setup hams my 14900KS. I’m not really sure why that is. But I certainly find it odd that these titles perform so well.

Just something I wanted to share.