Jump to content
HWBOT Community Forums

8500G / 8300G 2/1 Core globals?


Recommended Posts

Could this perhaps be a good moment to pause and consider having all the exotic categoriy splits in the first place?

We now have 3 types of cores per CPU, each available in pretty much single digit increments up to 16. And this will only get worse over time.

As a spectator, I don't give a rat's ass about, say, 14900K benched as 8P. It's sold as a 24c CPU, it's supposed to be used as a 24c CPU, ffs.

  • Like 5
  • Thanks 5
  • Confused 1
Link to comment
Share on other sites

19 hours ago, TaPaKaH said:

Could this perhaps be a good moment to pause and consider having all the exotic categoriy splits in the first place?

We now have 3 types of cores per CPU, each available in pretty much single digit increments up to 16. And this will only get worse over time.

As a spectator, I don't give a rat's ass about, say, 14900K benched as 8P. It's sold as a 24c CPU, it's supposed to be used as a 24c CPU, ffs.

I agree. From what I remember the initial argument was that, given your example, a real 24 core CPU like the Threadripper 7960X or the Xeon w7 2495X is always going to outperform the 14900K, since the 14900K is basically an 8c/16t with ecores. The solution that we have kinda works for now, but what happens once there are 3 different types of cores in the CPU? Are we really going to make 3 different categories for the same CPU? Or even more in the future? 

 

Edited by George_o/c
  • Like 1
  • Confused 1
Link to comment
Share on other sites

32 minutes ago, ale belo said:

on techpowerup it says that the difference is only on the cache so I assume that the cpu is 4c/8t

Exactly. In theory the cpu should be seen as a classic 4c/8t. The Zen4 core has "full" l3 cache, while the 3 Zen4c cores should have half l3 cache. Then it will be necessary to evaluate how much this impacts performance both at stock and in overclock, but also looking at the specifications on the AMD website, the CPU should be seen as a classic 4c/8t without any distinction between hypothetical core performance and core efficiency

  • Like 1
Link to comment
Share on other sites

1 minute ago, Zombie568 said:

OK, we have a fx 6300 that has 6 cores, of which 3 are full-fledged and 3 are “virtual”, why then is there no 3p category for it, but at the same time for older processors there is a 4p category

fear of losing globals?

Link to comment
Share on other sites

For example

Fx4100 2p- no category- no sub

Fx 6300 3p no category - no sub

Logic 100% can make these categories and the results will appear?

10 hours ago, Leeghoofd said:

FX-8120(4P) : 14 subs

FX-8320(4P) : 28 subs

FX-8350(4P) : 26 subs

FX-9590(4p) : 31 subs

and this even after including them in a competition.

 

 

Link to comment
Share on other sites

What the low-end fx series processors have  what the low-end fx series processors don't  the fact that the low-end FX series processors do not have their own 2p, 3p categories seems unfair to me. Why is that?

This is one of the series "this is allowed, and this is not allowed, because we are afraid that it may break some kind of balance, or simply someone will be dissatisfied"

Link to comment
Share on other sites

Zen 4 and Zen 4c core having same performance per clock, but different max clock and power efficiency. So 1 Core category should be allowed as we speak earlier on this forum before Alderlake released. We already discussing this. Btw Zen 4c core is already on AMD Ryzen Z1 ;)

Link to comment
Share on other sites

  • Crew

The FX (4P) categories were a test to see how popular one core per computing unit would be, now as you can see from the previous highlighted numbers, it is far from.

It has several reasons I guess:

  • Performance of this gen of AMD CPUs was a turndown
  • When they were released HWBOT didn't support any core disabling at all

No clue what the addition of these would cause any "imbalance", what take 3 core globals from another AMD CPU?

We just don't prefer to create new categories and spot that their hardware pages stay nearly empty...
 

Link to comment
Share on other sites

a strange decision for “testing” to open only the 4p category fx obviously loses in it. but if they released the same 3p, then fx would be much interesting in terms of global points

PS: I understand that the sponsor of the bot is Intel, but this is openly discrediting AMD processors...

 

8 minutes ago, Leeghoofd said:

The FX (4P) categories were a test to see how popular one core per computing unit would be, now as you can see from the previous highlighted numbers, it is far from.

It has several reasons I guess:

  • Performance of this gen of AMD CPUs was a turndown
  • When they were released HWBOT didn't support any core disabling at all

No clue what the addition of these would cause any "imbalance", what take 3 core globals from another AMD CPU?

We just don't prefer to create new categories and spot that their hardware pages stay nearly empty...
 

 

Link to comment
Share on other sites

  • Crew

Just from technical point on Phoenix2:

The little cores part in the concept is based on the same architecture: Zen4
A die shot of a Phoenix2 processor shows, that dies between Zen4 and Zen4c are different in size (almost twice) but the difference between them is only in some cache and other electrical parts.

image.png.0c6f897ea1de2ceb39033f9c38e99cae.png

One effect of this reduction in parts is also reduced frequencies of this cores. Now the question is, how much is this difference in an overclocked status. Means: Could these E-Cores possibly reach P-Cores clock speed? In default config we have here 1300 MHz difference between the highest clock status. However, because of the same architecture and same size of cache, the performance per clock between E and P Core should be nearly 1:1.

This is a difference to Intel where we really had different architecture between the cores. E-Cores where really designed to safe energy and are much weaker, in AMDs concept however, it looks more like a method to reduce costs. At least its my theory they wanted to build overall a small CPU and give them one or two powerful cores to let them not appear to be too weak. Thats another difference to Intel: E-Cores are more than P-Cores.

Now what does this mean for our DB:

I can only give my personal opinion here. The logic for splitting categories was always complex and never written down. In the past we had die shrinks for GPUs which could led to possibly higher clocks and made no split, before we had newer CPU revisions which added features like x64 and made again no split. Now creating 1P or 2P categories for those CPUs feels wrong here. Previously we always said, as an overclockers place, we do not care to much about to be a hardware museum and forcing 120 % correctness. Means: You have to use the best silicon available. If you have some old revision of hardware, its your bad if you can compete with it in the rankings. This avoids a state of making our database too complex. We keep it simple. And thats desperately needed because we already have thousands of categories. If AMD says these boys are 4 core or 6 CPUs than thats the way the meant to be benched. Again, we dont have here a one core CPU with 3 helper cores. We have a 4 core CPU with 3 cores which run a little bi slower and one core running with full frequency.

One last word to our submission code on our website:

The state of adding performance core categories besides the original hardware is quite nasty. We should already have the code in place to do this all within the same category because of the unlocking core feature introduced for AMD CPUs. Of course the logic is reverse here as we deactivate some cores but we already define how much p cores we have. Why is it so hard to adapt the code here? If we create p-core categories now for every new CPU, it will be much work to make things look clean in the future.

  • Like 2
Link to comment
Share on other sites

  • 2 months later...
On 1/11/2024 at 6:40 PM, chispy said:

It would be nice to move on from dinasour Phenom 550 unlock to 3 cores to this new 8300g 3 unlock to core aproach. In my opinion it is a huge step forward as we cannot get stuck in the past.

Hey, AM3 is still not as ancient as LGA775 for singlecore ;) Also easier to bin.

Edited by unityofsaints
  • Like 1
Link to comment
Share on other sites

Join the conversation

You can post now and register later. If you have an account, sign in now to post with your account.

Guest
Reply to this topic...

×   Pasted as rich text.   Paste as plain text instead

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.

×
×
  • Create New...