• 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 .
Robert896r1 said:
Something to note for those who're doing a lot of tuning on the core side. You can easily get into clock stretch scenarios and think you're doing a good job. Jufes/framechaers being a prime example.

Here he's convinced he's tuned x3d well above stock but pay attention to "core effective clocks" maximum figure. That never went past 4,745. Core effective max is the highest frequency that any core actually achieved. This has been an amd thing for quite some time and now i'm seeing in on the intel side as well, when people start messing with V/F curves which is intel's CO. Anyway, learn to read a few key metrics in HWINFO when tuning.
Click to expand...
Good catch.

I buy 7800x3d at some point for an upgrade for the added gaming experience it provides.
No real reason to upgrade from current but that x3d is nice
 
Robert896r1 said:
Something to note for those who're doing a lot of tuning on the core side. You can easily get into clock stretch scenarios and think you're doing a good job. Jufes/framechaers being a prime example.

Here he's convinced he's tuned x3d well above stock but pay attention to "core effective clocks" maximum figure. That never went past 4,745. Core effective max is the highest frequency that any core actually achieved. This has been an amd thing for quite some time and now i'm seeing in on the intel side as well, when people start messing with V/F curves which is intel's CO. Anyway, learn to read a few key metrics in HWINFO when tuning.

image.png
Click to expand...
Even worse is the fact that people are paying him $500 to tune their system for them.
 
  • Like
Reactions: J.D
Robert896r1 said:
Something to note for those who're doing a lot of tuning on the core side. You can easily get into clock stretch scenarios and think you're doing a good job. Jufes/framechaers being a prime example.

Here he's convinced he's tuned x3d well above stock but pay attention to "core effective clocks" maximum figure. That never went past 4,745. Core effective max is the highest frequency that any core actually achieved. This has been an amd thing for quite some time and now i'm seeing in on the intel side as well, when people start messing with V/F curves which is intel's CO. Anyway, learn to read a few key metrics in HWINFO when tuning.

image.png
Click to expand...

This "Core Effective Clocks" makes no sense to me, and i get that its both AMD and Intel now.

For my CPU at least, and i'm 100% sure it will be the same with Intel, the performance increase is pretty much 1 to 1 in line with clock increases i set, when i run clocks set for 4.9Ghz in R23, up from 4.6Ghz (a 300Mhz increase) i'm getting 6% more performance, perhaps slightly over, 6.2%, the percentage difference in clocks is 6.5%.

HWINFO says my "effective clocks" have only changed by about 100Mhz, erm? No.... BS! I think Ryzen Master is the only tool i trust for my CPU right now, it seems to be the only thing telling me the same thing i experience.
 
Last edited:
I think maybe HWINFO have been a bit annoyed since AMD introduced their own free windows environment hardware monitoring / editing tool along with Ryzen, and its pretty damned good.

I think Intel have one now too don't they?
 
humbug said:
This "Core Effective Clocks" makes no sense to me, and i get that its both AMD and Intel now.

For my CPU at least, and i'm 100% sure it will be the same with Intel, the performance increase is pretty much 1 to 1 in line with clock increases i set, when i run clocks set for 4.9Ghz in R23, up from 4.6Ghz (a 300Mhz increase) i'm getting 6% more performance, perhaps slightly over, 6.2%, the percentage difference in clocks is 6.5%.

HWINFO says my "effective clocks" have only changed by about 100Mhz, erm? No.... BS! I think Ryzen Master is the only tool i trust for my CPU right now, it seems to be the only thing telling me the same thing i experience.
Click to expand...
I mean what can you really trust to give the most accurate readings? I’ve had instances where HWInfo has been a little strange. The best thing for CPU and memory overclocking is to not just look at the numbers increasing but to check performance is actually getting better. With memory especially I’ve had times when the numbers are getting better but performance is going backwards. Clock stretching is a real thing but it’s a shame it didn’t always happen instead of getting a BSOD.
 
I've got a 7950x installed in an incomplete project I'm waiting on the final parts for, it's predominantly a gaming build. How big if a difference would I see between the 7950x and 7950x 3d?
 
Harshell123 said:
I mean what can you really trust to give the most accurate readings? I’ve had instances where HWInfo has been a little strange. The best thing for CPU and memory overclocking is to not just look at the numbers increasing but to check performance is actually getting better. With memory especially I’ve had times when the numbers are getting better but performance is going backwards. Clock stretching is a real thing but it’s a shame it didn’t always happen instead of getting a BSOD.
Click to expand...

I know clock stretching is a real thing, its a phenomenon that's been going on for decades, i know :)

But it is something where your performance does not increase in line with your clock increases, My Ryzen 1600 was guilty of clock stretching, it would stop scaling at about 3.9Ghz, i could get it as hight as 4.2Ghz but it would make no difference to performance if not even slightly regress.
But if you run 6% higher clock and you get 6% higher performance that is not clock stretching, HWINFO is telling me it is. So i don't trust it, its wrong. I performance test all of my settings, GPU's can do it too, on the core and the memory.
 
Vidar said:
I've got a 7950x installed in an incomplete project I'm waiting on the final parts for, it's predominantly a gaming build. How big if a difference would I see between the 7950x and 7950x 3d?
Click to expand...

Depends on what you're playing, anything from a slight regression because of slightly lower clocks to over 30% at the other end of the scale.
 
@Vidar not all games benefit from the 3D cache, older games generally don't, newer games, generally, do, on top of that by how much also depends on the game.
 
Last edited:
Vidar said:
720p because I've already spent over 4k on my system and will continue to use the same Dell special I've had for 20 years.

Or could it be because I want the absolute best system I can afford?
Click to expand...

Look at the res you play and games and compare both if it's worth it or not, guess you would need to sell the 7950x for a loss ?
 
henry hoov3r said:
I watched that earlier. He gets some odd results with the 7950x3d compared to the 5800x3d.
Click to expand...
What he doesn't realise is that by default COD will set the render worker count to almost match (it uses 1 below) how many physical cores you have, so this can put CPUs with a larger physical core count at a disadvantage, aka the 13900K/7950X3D et al.

COD works much better with a render count 6, or 7. Guess what the 5800X3D uses by default in COD? 7. This probably goes part way to explaining his odd result.

So, change the worker count thread to 6, watch the 7950X3D increase performance significantly that nothing else will get near it.


Intel CPUs would also benefit from this too.
 
Last edited:
You must log in or register to reply here.
Back
Top Bottom