Caporegime
Furry muff, core cycler I have more faith in so yeah maybe a bad one then. Good luck with the RMA.
-
Competitor rules
Please remember that any mention of competitors, hinting at competitors or offering to provide details of competitors will result in an account suspension. The full rules can be found under the 'Terms and Rules' link in the bottom right corner of your screen. Just don't mention competitors in any way, shape or form and you'll be OK.
Poll: Ryzen 7950X3D, 7900X3D, 7800X3D
- Thread starter kindai
- Start date
Associate
Wish I could live with adding a positive curve but that will hurt resale value and it is faulty by nature of how it should operate at stock.
Just hoping the replacement has a good IMC. A drop in SP rating, fine, but I'll be upset if FCLK/IMC sucks
This chip has it all seemingly apart from one dodgy core. A core which is rated the best on the CCD to boot 
Last edited:
Caporegime
Yeah understandable. I had to return one 7950X3D to CeX for a similar reason. That one would randomly restart in Windows without +8 CO on the two strongest cores. It was a decent sample overall on benchmark scores and did 2200 FCLK. Was also a SP112 batch.Wish I could live with adding a positive curve but that will hurt resale value and it is faulty by nature of how it should operate at stock.
Just hoping the replacement has a good IMC. A drop in SP rating, fine, but I'll be upset if FCLK/IMC sucks
Associate
Yikes, someone over on OC.net also had this issue of needing a positive curve on the best core. That's 3 of us now. Small sample size, sure, but that really looks dodgy for AMD lol.
8 and 9 are my two strongest cores
(Ryzen Master +1 in naming as it doesn't start at 0)
Was your issue after the chips frying themselves drama? I just joined AM5 but I wonder if some of these chips were binned really close to the edge and whatever changes AMD has made to AGESA since that drama has resulted in this? No idea, just speculating how chips make it through Q&A that literally have cores assigned the best fail at stock settings.
Last edited:
Caporegime
I only had it on one sample and I’ve tried 8 or so (in total).Yikes, someone over on OC.net also had this issue of needing a positive curve on the best core. That's 3 of us now. Small sample size, sure, but that really looks dodgy for AMD lol.
8 and 9 are my two strongest cores
(Ryzen Master +1 in naming as it doesn't start at 0)
Was your issue after the chips frying themselves drama? I just joined AM5 but I wonder if some of these chips were binned really close to the edge and whatever changes AMD has made to AGESA since that drama has resulted in this? No idea, just speculating how chips make it through Q&A that literally have cores assigned the best fail at stock settings.
Last edited:
Yikes, someone over on OC.net also had this issue of needing a positive curve on the best core. That's 3 of us now. Small sample size, sure, but that really looks dodgy for AMD lol.
8 and 9 are my two strongest cores
(Ryzen Master +1 in naming as it doesn't start at 0)
Was your issue after the chips frying themselves drama? I just joined AM5 but I wonder if some of these chips were binned really close to the edge and whatever changes AMD has made to AGESA since that drama has resulted in this? No idea, just speculating how chips make it through Q&A that literally have cores assigned the best fail at stock settings.
It was never an issue.
Associate
The VSOC/ASUS thing? Given AMD ended up restricting VSOC voltage in the BIOS I'd say it was.
Or not passing AVX2 stability tests with stock settings? I've just gone down that rabbit hole and sadly all throughout 2023 people were having issues let alone me with a dud chip in 2024
AMD QA seems absolutely awful on 7950x/7950x3D chips. I'd advise everyone checks if their CPUs can run AVX2 instructions at stock.
Having to turn on PBO, which is technically out of spec, to add a positive curve value to a core just to pass stability checks is unacceptable and shows a faulty product. Especially when other chips pass fine at stock values.
Last edited:
Associate
Confirmation from advice online that if you find the problem cores and turn on PBO and add a positive offset it might work. Both core 8 & 9 need a positive value to pass stability.
Which is obviously unacceptable.
Sent off a ticket last night to OCUK for an exchange.
edit - RMA accepted already, OCUK really on the ball these days with CS.
Last edited:
Yeah.
Associate
DPD notification already for replacement parcel, that's some proper good customer service from OCUK!
I’m going to be assembling a 7950X3D system in the next few days, what do I need to run to check this?
Soldato
I had a 7900X3D in my PC, its now in my Mrs machine, I think unless you arte gaming all the time they are a load of rubbish, i'll never buy a 3D CPU again, this is my 7900 non X in my machine, its currently stable in every test ive thrown at it, and theres a lot of them, RealBench, Prime 95 Small FFT's AVX and non AVX, Prime 95 Large FFT AVA nad non AVX, Core Cycler, Cinebench, YCruncher, OCCT all sorts, its +150mhz boost and currently -15 CO on all cores, I will be pushing it further yet, this has got to be better than a golden sample.
Last edited:
Associate
- Joined
- 30 Mar 2017
- Posts
- 857
Nice results, especially with the 150mhz boost! That's perfect. I haven't touched that setting yet. Shall I dabble?I had a 7900X3D in my PC, its now in my Mrs machine, I think unless you arte gaming all the time they are a load of rubbish, i'll never buy a 3D CPU again, this is my 7900 non X in my machine, its currently stable in every test ive thrown at it, and theres a lot of them, RealBench, Prime 95 Small FFT's AVX and non AVX, Prime 95 Large FFT AVA nad non AVX, Core Cycler, Cinebench, YCruncher, OCCT all sorts, its +150mhz boost and currently -15 CO on all cores, I will be pushing it further yet, this has got to be better than a golden sample.
I have the 7900X and from all the issues it seems others have had simply getting the X3D to run correctly, doesn't seem massively worth it. Perhaps when it is perfected in both hardware and software in the future it will be more viable.
My chip does -30 CO on all cores, anything above and it just starts to throw errors.
Sometimes the 'best' chip such as X3D isn't always the best I have learnt.
Loving the slight power reduction and slightly lower temps as well.
Honestly feel the Vanilla 7000 series non X was the option for this current Gen.
Interested to see what 8000/9000 has to say about that!
Last edited:
Soldato
Yes dabble away, free extra performance is always worth it, you'll probably find you'll have to reduce your CO values though.
The one that will get my real attention is when they move the 3D cache to the IO die so that all cores can use it, they could probably considerably bump the clock speeds up too then on the chiplets as they are limited currently due to the 3D cache being directly on the chiplet, I thought it was coming in 9000 series, but it seems not, maybe Zen6, but I think that will be the last AM5 CPU we will see for the AM5 platform.
Associate
Download OCCT and run the extreme variable large test. Put it on Corecycler.
Thanks, let us know how you get on with the replacement. I hope it’s not a case of “they’re all like that sir” and OcUk reject your RMA.
Soldato
Delete
Last edited:
Yes/no - it is possible to avoid any scheduling issues with the CPU, but it isn't always as trivial as some like to make out.
Soldato
The only time I've had a issue with a game not going onto the 3d chiplet was the Once Human alpha as game bar didn't detect it automatically, ticked the remember this is a game and was all ok.