• 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

Will you be purchasing the 7800X3D on the 6th?

  • Total voters
    191
  • Poll closed .
iron_monkey said:
I meant in general for optimising ram and cpu, ivve got hwmonitor,aida64, cinebench23 and zentimings so far
Click to expand...
For testing system stability I use the following:

I use y-cruncher (options 1-7-0 + enter when launched) for testing CPU stability when using curve optimiser. Leave for a minimum of 8-12 hours. If things pass then you can change settings and retest. Use 100% CPU fan speed, this test will create a lot of heat for the CPU. The varied workloads of y cruncher will usually spot a core that has too much curve optimiser applied and if it finds an unstable core it will stop the test and tell you. You can then enter the BIOS and reduce the curve optimiser for that core from say -20 to -15 and retest. It takes some time but is worth it in the end. I took about two weeks or more to fully tune my CPU and cores down to the point that two notches up on any of my 16 cores will fail y cruncher.
I use RAM Test (Karhu) (Under Advanced set CPU Cache to enabled) for testing memory frequency/timings and FCLK stability. Minimum of 8-12 hours testing required to properly test memory and FCLK. I recommend a minimum of 10 hours.
This chart (look for Karhu heading) will give you an indication of what MBs you should get for the memory frequency/timings used. If you look at the screenshot I posted above, my Karhu MB numbers line up with expected performance. If your FCLK is too high you can see a drop of 50-100MBs of speed in RAM Test (Karhu). Hover over the % sign in Karhu to check your speed. Generally it takes 30-60 minutes before the numbers stabilise so don't pay too much attention to the numbers for the first 5-10 minutes.
Zen timings is useful for checking on voltages applied, VDDP, SOC etc and if you have geardown mode disabled. If you join the Zen Timings Discord you can get the latest beta version.
Cinebench is completely useless for everything so don't bother with it.
 
LtMatt said:
For testing system stability I use the following:

I use y-cruncher (options 1-7-0 + enter when launched) for testing CPU stability when using curve optimiser. Leave for a minimum of 8-12 hours. If things pass then you can change settings and retest. Use 100% CPU fan speed, this test will create a lot of heat for the CPU. The varied workloads of y cruncher will usually spot a core that has too much curve optimiser applied and if it finds an unstable core it will stop the test and tell you. You can then enter the BIOS and reduce the curve optimiser for that core from say -20 to -15 and retest. It takes some time but is worth it in the end. I took about two weeks or more to fully tune my CPU and cores down to the point that two notches up on any of my 16 cores will fail y cruncher.
I use RAM Test (Karhu) (Under Advanced set CPU Cache to enabled) for testing memory frequency/timings and FCLK stability. Minimum of 8-12 hours testing required to properly test memory and FCLK. I recommend a minimum of 10 hours.
This chart (look for Karhu heading) will give you an indication of what MBs you should get for the memory frequency/timings used. If you look at the screenshot I posted above, my Karhu MB numbers line up with expected performance. If your FCLK is too high you can see a drop of 50-100MBs of speed in RAM Test (Karhu). Hover over the % sign in Karhu to check your speed. Generally it takes 30-60 minutes before the numbers stabilise so don't pay too much attention to the numbers for the first 5-10 minutes.
Zen timings is useful for checking on voltages applied, VDDP, SOC etc and if you have geardown mode disabled. If you join the Zen Timings Discord you can get the latest beta version.
Cinebench is completely useless for everything so don't bother with it.
Click to expand...
Thanks Matt, you really are a helpful chap. When i get a chance i'll try and find a guide for my board. :D
 
DeNiM said:
Hello everyone, hope everyone is doing well and wish everyone a nice weekend.

I am very happy that I found this forum and hope that someone can help me. I have to say that I am not familiar with PCs & Co. at all.

I had a completely new PC built a few months ago. But unfortunately I'm very disappointed because it's not going really well. I'm currently only playing CoD MWII/WZ2 and I'm very dissatisfied with the FPS. And for that reason I hope someone can help me via discord or something like that to make my pc run really well. I put 2 pictures of the CoD benchmark in.

This is installed in my PC:

AMD Ryzen 9 7950X3D
ASUS ROG Strix RTX4090 OC
ASUS ROG Strix Helios
ASUS ROG Ryujin III 360
ASUS ROG Thor 1200P2 Gaming
ASUS ROG CROSSHAIR X670E Extreme
4x 16GB Corsair Dominator Platinum CMT32GX5M2X5600C36
2x Samsung 990 Pro 2TB
9x Corsair QL120
1x Corsair QL140 RGB
Windows 11 Pro

I thank you in advance for the help and stay healthy everyone.

Best regards from Germany

DeNiM

IMG-4368.jpg

IMG-4367.jpg
Click to expand...
Your cpu is running fine. Don’t take much notice to others results that are much higher.

This is my result without ram tuning.

https://imgur.com/a/OJPYvyY

I would get exactly the same as your result without certain tricks to get higher numbers. It have nothing to do with your cpu running badly.

Certain things you can install to get a good bump in numbers. Things you can enable/disable etc.

And this result I’m getting is still low compared to Beefy Tech for example.

You will find people trying to make things look better than they are for 99.9% of users because without this the CPU won’t look as good, and many would not buy.

Also Y Cruncher for Curve Optimiser is not so great. Nothing works great. You can get leak from one core to the next which can greatly affect the core or the next core being unstable. And don’t forget to test idle stability because you can be YCruncher 12 hours stable and crash at desktop doing nothing. There is no test for this idle stability either.

Your ram tuning can affect your Curve Optimiser stability and your Curve optimiser can affect your ram stability.

All in all I think this CPU is a nightmare to tune for very little gains. Unless you using tricks.

Unless these people do proper testing openly on a system that is set up with just a clean install of windows and no tricks I would not believe anything.

I have tested this myself and have found so many people asking “why is mine not performing correctly?” The answer is simply yours is performing correctly. You have been looking at people results with stripped windows and countless other tricks to get higher numbers. What the vast majority will get is much lower when you use everyday windows snd 0 other tricks including modded power plans.

My only other thing I would say is to not listen about this process lasso for 7950x3d. It does not make performance better than using the correct way.
 
LtMatt said:
Yeah it's something like that. Well worth the money though if you want to properly tune your system.

Might want to have a fan pointed at your memory whilst running it with voltage over 1.4v though as they sticks will get warm.
Click to expand...
Cheers My case has 3 x 140 front fans so hopefully will be ok im not going crazy overclocks id just like to tighten timing if possible, is there an easy way to tell if i have A die?

VvGEhNN.png
 
Last edited:
Brizzles said:
A Die is usually 6800 +.

You'll not find A die on 6000 kits - unless they advertise such ( or very very lucky )

A die ends in A on the module sticker ;

2j2xuPi.png
Click to expand...
Not strictly true, it uses a die on some 6000 SR and DR kits. The first 32GB SR kit I ohad GSkill EXPO C30 was a die. Likewise I had two similar GSkill 6000C30 kits but DR and they were both a die too. M die more common though tbf.
 
Last edited:
McCarthy said:
Your cpu is running fine. Don’t take much notice to others results that are much higher.

This is my result without ram tuning.

https://imgur.com/a/OJPYvyY

I would get exactly the same as your result without certain tricks to get higher numbers. It have nothing to do with your cpu running badly.

Certain things you can install to get a good bump in numbers. Things you can enable/disable etc.

And this result I’m getting is still low compared to Beefy Tech for example.

You will find people trying to make things look better than they are for 99.9% of users because without this the CPU won’t look as good, and many would not buy.

Also Y Cruncher for Curve Optimiser is not so great. Nothing works great. You can get leak from one core to the next which can greatly affect the core or the next core being unstable. And don’t forget to test idle stability because you can be YCruncher 12 hours stable and crash at desktop doing nothing. There is no test for this idle stability either.

Your ram tuning can affect your Curve Optimiser stability and your Curve optimiser can affect your ram stability.

All in all I think this CPU is a nightmare to tune for very little gains. Unless you using tricks.

Unless these people do proper testing openly on a system that is set up with just a clean install of windows and no tricks I would not believe anything.

I have tested this myself and have found so many people asking “why is mine not performing correctly?” The answer is simply yours is performing correctly. You have been looking at people results with stripped windows and countless other tricks to get higher numbers. What the vast majority will get is much lower when you use everyday windows snd 0 other tricks including modded power plans.

My only other thing I would say is to not listen about this process lasso for 7950x3d. It does not make performance better than using the correct way.
Click to expand...

Going back to when Matt replied to you here....

Ryzen 7950X3D, 7900X3D, 7800X3D

Just bought the 7950x3d. What’s the best way to run this for gaming? I’ve seen people say that Process Lasso is the best way, but surely that’s not what AMD recommend. Would installing all the drivers as per AMD’s recommendations work just as well? I have watched some YouTube videos but it seems...
forums.overclockers.co.uk forums.overclockers.co.uk

.....has anything changed much for you with your thoughts when you first bought a 7950x3D....? Not sure if this was your first AMD CPU but are you overall happy with it for performance and glad you bought that one over the 7800x3D..?

So a typical out of box experience, with game bar, is going to get the most performance from the 7950x3D without really working hard to achieve anything more..?

Thanks.
 
You must log in or register to reply here.
Back
Top Bottom