AMD Ryzen 9 5950X, 5900X Geekbench results unearthed

[HEXUS]

And the first image of these desktop processors 'in the flesh' is shared online.

Read more.

[will19565]

i wish Intel, AMD and Nvidia would all release their CPU's and GPU's at the same time, it would make comparisons sooooo much easier.

[LSG501]

Not that I'm a huge fan of geekbench but including the 3900x/3950x would have been nice, and arguably more relevant for some users.

3950x
1292 - Single-Core Score
14097 - Multi-Core Score

3900x
1280 - Single-Core Score
11861 - Multi-Core Score

That kind of brings out a bit of an anomaly on the multicore for the 5950x, the odds of it being slower than the 3950x is pretty slim considering all other 'benches' we've seen....

[Corky34]

i wish Intel, AMD and Nvidia would all release their CPU's and GPU's at the same time, it would make comparisons sooooo much easier.

Every reviewer on the planet are sharpening their pitchforks for you now.

Seriously though it would be a mess if everyone released all their CPU's and GPU's at the same time, can you imagine trying to review upwards of 10+ different products all on the same day.

[ik9000]

Every reviewer on the planet are sharpening their pitchforks for you now.

Seriously though it would be a mess if everyone released all their CPU's and GPU's at the same time, can you imagine trying to review upwards of 10+ different products all on the same day.

it would show who were the paid reviewers but apart from that, chaos!

[ohmaheid]

Maybe the 5xxx series were being run at lower clocks?

(In reply to LSG501)

[ohmaheid]

Not that I'm a huge fan of geekbench but including the 3900x/3950x would have been nice, and arguably more relevant for some users.

3950x
1292 - Single-Core Score
14097 - Multi-Core Score

3900x
1280 - Single-Core Score
11861 - Multi-Core Score

That kind of brings out a bit of an anomaly on the multicore for the 5950x, the odds of it being slower than the 3950x is pretty slim considering all other 'benches' we've seen....

Maybe the 5xxx series were being run at lower clocks?

[LSG501]

Maybe the 5xxx series were being run at lower clocks?

Possible but whatever it is it's lower than it should be imo, AMD scales pretty well with core count (always have done in all honesty) and it seems to do it quite well with the 5900x.

It just doesn't seem 'right' for the newest cores which are faster per clock to be slower, even with a 100mhz lower clock (which the 5900x has as well), in multi-core score on the 5950x. In all honesty I'd expect closer to the 15,250 range for the 5950x based on some rudimentary assumptions and maths based off the 3xxx series.

[Corky34]

Possible but whatever it is it's lower than it should be imo, AMD scales pretty well with core count (always have done in all honesty) and it seems to do it quite well with the 5900x.

It just doesn't seem 'right' for the newest cores which are faster per clock to be slower, even with a 100mhz lower clock (which the 5900x has as well), in multi-core score on the 5950x. In all honesty I'd expect closer to the 15,250 range for the 5950x based on some rudimentary assumptions and maths based off the 3xxx series.

Being bored i started browsing geekbech multi-core scores for the 3950x and they're all over the place, I'm guessing that 14097 - Multi-Core Score for the 3950x must be an average or something as scores seem to jump around from the high 800's on linux to the high 1600's on OSX and Windows.

[LSG501]

Being bored i started browsing geekbech multi-core scores for the 3950x and they're all over the place, I'm guessing that 14097 - Multi-Core Score for the 3950x must be an average or something as scores seem to jump around from the high 800's on linux to the high 1600's on OSX and Windows.

I just took the first link that came up in google tbh, but those varaitions could just be geekbench being geekbench.... It's often used to compare stuff between operating systems/architectures but they're all compiled differently and make use of any coprocessors etc so the only real use it has is comparing like for like and even then I wouldn't say it's necessarily reliable.

[Corky34]

just be geekbench being geekbench....

Yea that's the point i was trying to make, sorry that it came across as questioning the score you posted as that wasn't the intention, tis why i said geekbench must be basing that on an average or something.

Either way TIL geekbench is strange and doesn't seem to be much use as a benchmark.

[LSG501]

doesn't seem to be much use as a benchmark.

In all honesty it isn't really unless you're an apple fan and want to quote the high numbers it gets to 'support' your choice of an iPhone lol