News AMD Ryzen 9 3950X Beats $2,000 Intel Core i9-9980XE In Geekbench Test

[Admin]

Leaked Geekbench benchmark puts AMD's recently announced 16-core Ryzen 9 3950X above the $2,000 18-core Intel Core i9-9980XE

AMD Ryzen 9 3950X Beats $2,000 Intel Core i9-9980XE In Geekbench Test : Read more

 

[JamesSneed]

If this is true then Intel is in for a hard 18 months in the HEDT and server space. AMD's efficiency is only going to prove out to be more important on really high core count chips.

 

[penn919]

If this is true then Intel is in for a hard 18 months in the HEDT and server space. AMD's efficiency is only going to prove out to be more important on really high core count chips.

I could believe that an R9 3950x could beat an i9-9960X...and that would be impressive in its own right, but there's no way it could beat a i9-9980XE. It has 18 cores with HT and Zen 2 is (allegedly) only marginally above/below skylake.

Intel would be shitting bricks if this were actually true though. It would mean AMD undercut their HEDT higher core count cpu by well over half the price...with better performance!

 

[mdd1963]

"...Apparently, AMD adidn’t test the new chip on the latest ....."

Where's that Mr. Paperclip/Clippy feller when you need him...???

 

[mdd1963]

I'm not sure how well Geekbench really measures true performance, or to what degree it might be awarding points for clockspeed, cores, assorted cache values, RAM speed, etc; the CInebench and Blender victories by 3900X are much more tangible, at least to me...

 

[joeblowsmynose]

I could believe that an R9 3950x could beat an i9-9960X...and that would be impressive in its own right, but there's no way it could beat a i9-9980XE. It has 18 cores with HT and Zen 2 is (allegedly) only marginally above/below skylake.

Intel would be shitting bricks if this were actually true though. It would mean AMD undercut their HEDT higher core count cpu by well over half the price...with better performance!

Well, the $499 12 core 3900x appears (wait for 3rd party reviews!) to already be mopping the floor with the $1200 Intel HEDT 12 core ... I wonder how it fares against their 14 core?

 

[joeblowsmynose]

I'm not sure how well Geekbench really measures true performance, or to what degree it might be awarding points for clockspeed, cores, assorted cache values, RAM speed, etc; the CInebench and Blender victories by 3900X are much more tangible, at least to me...

Yeah geekbench is stupid, I don't know why people use it. It often makes weird results that make no sense at all. That said, it generally seems to heavily favour intel -- well at least before some of these new Zen2 benches ...

I know Intel fanboys are hating Cinebench these days because it makes AMD look good in multithreaded loads (which Ryzen easily dominates in per dollar multi performance), but the single threaded run aligns well with what you might expect from bottlenecked gaming performance differences. You'll be hearing much more crying from them about Cinebench now that Zen2 appears to be whooping Intel on both metrics, lol. (but wait for the 3rd party reviews people!)

The way I see it - if I have both the Cinebench single threaded and multi-threaded performance numbers, I pretty much got what I need - no need to look much further - I know pretty much how that will translate into the performance on the things I do.

I do a fair bit of 3D rendering and megatasking as well, so that Cinebench multi score is extremely relevant to me.

 

[SkyBill40]

I could believe that an R9 3950x could beat an i9-9960X...and that would be impressive in its own right, but there's no way it could beat a i9-9980XE. It has 18 cores with HT and Zen 2 is (allegedly) only marginally above/below skylake.

Intel would be shitting bricks if this were actually true though. It would mean AMD undercut their HEDT higher core count cpu by well over half the price...with better performance!

And if it does, in fact, turn out to beat the i9-9980XE... will you come back here and admit you were wrong? Or will you just ignore your comment having been made at all?

 

[bcroner]

To the editor (users): although it’s likely that some use will see a much bigger improvement than others. => although it’s likely that some users will see a much bigger improvement than others.

 

[penn919]

And if it does, in fact, turn out to beat the i9-9980XE... will you come back here and admit you were wrong? Or will you just ignore your comment having been made at all?

The math doesn't work out, so I can't be wrong...at least in general. Geekbench might be a poor bench that doesn't accurately depict cpu performance. Think about it, if IPC is roughly the same and the intel has more cores, then it couldn't possibly perform underneath the 3950x. Not in multithreaded workloads at least.

 

[SkyBill40]

The math doesn't work out, so I can't be wrong...at least in general. Geekbench might be a poor bench that doesn't accurately depict cpu performance. Think about it, if IPC is roughly the same and the intel has more cores, then it couldn't possibly perform underneath the 3950x. Not in multithreaded workloads at least.

Well see soon enough, won't we? And when we do, should the 3950X beat the 9980XE as has been mentioned, will you admit you were wrong? That is the real question here and you've failed to answer it despite it being a rather simplistic yes or no answer.