[SOLVED] 5600x lower than average performance for single core

[Cookiezaddict]

Hi guys,


I have just general question, out of curiousity, maybe it is not even an issue but well…. i will ask the questions anyway

my system are as follow;


CPU: Ryzen 5 5600X

GPU: TUF RX 6800 XT

Motherboard: Strix B550f GamingWifi2

PSU: 850w EVGA G5

CPU Cooler: Be Quiet! Pure Rock 2

Main Storage: 1 TB Samsung 980 M.2

Memory: 16GB 4000mhz cl16 Corsair LPX @ 3800mhz fclk mclk uclk 1:1:1 1900mhz.

Windows 11 Pro

The case is basically more towards single core performance, which could be normal, but slightly lower than what I expected.

So when running stock setting, cinebench r23 will score around 10.9-11.1k on multi core, 1490-1520 on single core. In this situation you can tell the MC result is normal, but the SC is lower than most 5600x

So move to PBO, no boost, ppt up to 100-120w doesn’t matter, tdc 75-90 and edc up towards 120-160, all will basically give me MC results around 11.9-12.1k for cinebench r23, even when CO only set at -5 all cores, no need to go extreme and for that? I am impressed. Without CO, it can only boost to 4.59ghz all cores and score will be slightly lower, around 11.5-11.7k on MC, again it’s normal normal, quiet good.

However, the Single core tho, if not boosted it will never go above 1540, if boosted 100mhz, it can reach 1570, if boosted 200mhz it can reach 1590-1600 on Cinebench R23.

Does anyone else think it’s a low Single Core performance for 5600x? Because online, I see the results of single core for this cpu would be 3-4% stronger than what I could get today. Multi core isn’t my concern here.

I tested on Geekbench, R20 Cinebench, all the same, my results is slightly lower than we can expect from 5600x

no thermal throttle, this set up only runs up to 84c on the most hardcore Prime95 even after 9 hours of stable test. Peak around 69-71’C on Cinebench r23 multicore, and only peak at around 51-53’C on single core test.

My timespy score for cpu test is around 8800 when boosted 100mhz, stays the same when boosted 200mhz, or no boost at all, because apparently it’s kind of multi core test

my 3dMark CPU profile test put me in great position for multi core results, and acceptable for single core result.

I would like to know, not about your 5600x score, but can you spot anything that may cause this?

Does anyone think it is because Win11? If it is win11, what can potentially cause this? Because even after fresh install, bios flash, I tried it all, all the same.

I do not turned my core integrity virtualisation no, so I am very curious what’s causing this

tried all LLC, tried vCore offset, tweak all I can possibly tweak changing global c state, and some things on bios, for weeks trying to get what most online reviewer claim, 1560-1590 single core R23 for 5600x, I can’t just can’t get it. Lol

thanks for checking and reading all the way.

hope I could get someone help to solve the mystery

 

[Cookiezaddict]

Let's make one thing clear. ANY and ALL "scores" that you are seeing where people have submitted the results of their benchmarks for the same CPU you have DO have PBO and boosting enabled. If you don't, you are never going to have the same scores, and again, even if you DO, you might not depending on how much memory you have installed, how fast your memory is, how tight your timings are, what motherboard you have (Yes, JUST the motherboard can have up to a 5% different in performance with all other things being equal), what other hardware you have installed and exactly what software you have installed. So, by all means, make every effort to eek out every ounce of performance you can for your own system, but base your results against your own system, not somebody else's. You'll just be trying to catch smoke.

Well basically all I feel is, does 5600x really do at least 1570 stock? Because I am not seeing forum for this, I am comparing to contents creators and articles that clearly stated STOCK vs PBO vs PBO boosted.

And I figured in average, single core stock stock, would be around 1550-1570, single core PBO could be 1570-1590, single Core PBO boosted will range between 1590-1640.

That is all,

and I am looking forwards to fellow 5600x users to confirm that, to share their single core performance, because me? I shared mine and it is 1520-1540 single no boost no PBO, 1540-1550ish PBO no boost, 1570 100mhz boost, 1590-1600 200mhz boost.

I am not trying to catch anyone here, but more like making sure I am not missing anything that could potentially affect performance, more like enjoying my LEARNING CURVE into gaming pc that I build without even being a gamer. Yes I don't really play games but enjoy understanding my system in a unique way, full of curiousity.

 

[Cookiezaddict]

Single core benchmarks will vary, depending on the instruction set used.
Such instructions may have little resemblance to your games or whatever the single thread performance is used to achieve.
Best to not obsess about it unless something is really off.

As another test, try the cpu-Z bench and look at the single thread performance rating.
It should be about 643:
https://valid.x86.fr/bench/rsf5p1/1

Again, realize that the cpu-z test submissions may reflect higher than normal performance.

Thanks for your advise man, yeah true, the obsession itself is not good I realised it too, it's like I am chasing 2% that I feel entitled LOL

This is my result, it's actually above what you mentioned? Maybe I am just overthinking afterall? I am running this with internet browsers full on, armory crate on for aura sync.

AMD Ryzen 5 5600X @ 4698.9 MHz - CPU-Z VALIDATOR (x86.fr)

 

[kerberos_20]

101mhz fsb can also raise cpu clock a little bit, i wouldnt go more than this, nvme doesnt like higher numbers, if no nvme, then around 110, which also helps GPU

see my signature, 3800X faster than 3800XT in both single thread/multithread in cpu-z
with just +1 mhz on fsb
multiplier x 1mhz = +44mhz which for 4.4GHz is 1% more clock, and benchs gets 10% more score...dunno why

 

[Karadjgne]

5600x default uses the full 65w.
5600x default freed of power constraints by PBO uses 74w. PBO does nothing to the cpu but allow it to use whatever power it wants to attain whatever boosts it can.

Ryzens boost according to 3x things. Voltages, loads and temps. Assuming that R23 puts the same load single core, and assuming that each cpu is using the same (or close enough) voltages, that leaves cooling as a wild card variable.

Default settings are all going to see the same R23 result of @ 1540, but with PBO enabled and power limits lifted, a person sitting in a 60°F room with a Fractal Torrent case, running max fans with a NH-D15 cpu cooler is going to see significantly different R23 results single core to someone in India, in the middle of summer with nothing but a ceiling fan, wide open windows and a Wraith Spire cpu cooler.

You are trying to compare apples and oranges and are acting miffed about getting grape juice. Those online results are only reflective of what's possible, not what's probable, there's simply too many variables to base an entire hypothesis of 'I must get this %' based on a synthetic benchmark on your pc, compared to others whom guaranteed have different equipment giving them a 3-4% advantage.

 

[CountMike]

Benchmarks are best for comparing "Before to after" HW, OS, driver or settings changes, Comparing to"Others" only in radical differences.