News AMD's Ryzen 9 5950X Shatters PassMark Records

The 2700x has been great to me but a RTX 3080 or (likely) Radeon 6900xt would certainly push it to and probably beyond it's limits. In RDR2 I have played on my system and on an Intel 10900k with an RTX 2080 and both with 3600mhz Corsair RAM.... you notice a big difference in the FPS on higher settings, so the 2700x is hitting it's ceiling on new GPU's.
 
  • Like
Reactions: Avro Arrow

VforV

Respectable
BANNED
Oct 9, 2019
578
287
2,270
We're not underestimating Zen 3, but it's a bit hard to swallow that the AMD chip with a 400 MHz boost clock would outperform the Core i9-10900K.
*400 MHz lower boost clock you mean.

Also, "a bit hard to swallow"? There are dozens of leaked benchmarks by now shopwing the same scenario, yet it's still hard to swallow for some people.... pffft. Well prepeare to "swallow" for real in 2 weeks time when AMD kicks Intel's a***...
 

Psyadin

Reputable
Jul 24, 2017
1
1
4,515
Would've thought TH would know about IPC, but apparently they dont, frequency is only one part of the equation, it would be like measuring your property size by length and no width.
In synthetical tests where more each cycle is more or less maxed out, the equation is frequency x ipc, to my knowledge Intel and AMD doesnt give an exact number, and it would be hard to verify even if they did, but we know for certain that AMD's is well over Intel's since at least Zen 2.
So the 400MHz difference doesnt matter that much, and assuming AMD's IPC is 10% or more over Intels it will far outweigh that difference... at least in synthetic benchmarks.
 
  • Like
Reactions: Avro Arrow
Well this is no surprise, it was kinda what you would expect after seen the Ryzen 5 5600X result last week..

I know this is only a leak, and only one test, but the most amazing part for me is that the Ryzen 5 5600X (6 cores / 12 threads) appears to have a multicore result of ~22800 points, when the Core i9 10900K have 24200 point with its 10 cores and 20 threads, that for me is kinda mind blowing.
 

semitope

Distinguished
Jun 2, 2006
35
4
18,535
Till actual benchmarks show up, these passmark results with 1 CPU of unknown settings should be ignored. Everything else you'll compare it to would have thousands of results averaged
 
The 2700x has been great to me but a RTX 3080 or (likely) Radeon 6900xt would certainly push it to and probably beyond it's limits. In RDR2 I have played on my system and on an Intel 10900k with an RTX 2080 and both with 3600mhz Corsair RAM.... you notice a big difference in the FPS on higher settings, so the 2700x is hitting it's ceiling on new GPU's.
Yes it is hitting the ceiling because the R7-2700X is on the Zen+ node which is just a smidgen better than the original R7-1700X. You'd get a massive performance boost just from going to the 3700X, let alone the 5700X.
Would've thought TH would know about IPC, but apparently they dont, frequency is only one part of the equation, it would be like measuring your property size by length and no width.
Well, lately, Tom's has been making some mistakes based on the outdated assumption that anything from AMD or ATi must be inferior to anything from Intel or nVidia unless proven otherwise. You should have seen Jarred's reaction when I speculated that AMD only teased the RX 5800 XT at the Ryzen 5000 launch instead of the RX 6900 XT. He called that idea "unrealistic" and it turned out to be correct. They have this idea that veteran users like me (who were here when Tom actually ran this site), are as clueless as someone who has been exposed to tech for two weeks.
Till actual benchmarks show up, these passmark results with 1 CPU of unknown settings should be ignored. Everything else you'll compare it to would have thousands of results averaged
They weren't ignored when the CPUs in question were Intel CPUs so they won't ignore them now.
 
  • Like
Reactions: Jim90

spongiemaster

Admirable
Dec 12, 2019
2,273
1,277
7,560
*400 MHz lower boost clock you mean.

Also, "a bit hard to swallow"? There are dozens of leaked benchmarks by now shopwing the same scenario, yet it's still hard to swallow for some people.... pffft. Well prepeare to "swallow" for real in 2 weeks time when AMD kicks Intel's a***...
It isn't surprising that the 5950x won, I think they're pointing out the margin of victory. The 5950x finished 16.3% ahead of the 10900k with a 12.25% clock deficit. That would indicate a nearly 30% higher IPC than the 10900k, which would be higher than even AMD claimed using the 19% better than the 3000 series claim as a basis. This is for single threaded performance. In multi-threaded, the 5950x is obviously going to obliterate the 10900k.
 
  • Like
Reactions: VforV and Briny

Zeh

Distinguished
Dec 7, 2010
169
0
18,690
What I find most interesting, is that the 5600X ST beats 10900K while almost beating it on multicore, even having only 6 vs 10 cores. It seems like a thermal limit on the 10900K. The 5800X will probably beat 10900K on both ST/MT. There's not much increase in MT for 5950X, perhaps also a thermal limit?
 

Makaveli

Splendid
He called that idea "unrealistic" and it turned out to be correct. They have this idea that veteran users like me (who were here when Tom actually ran this site), are as clueless as someone who has been exposed to tech for two weeks.

Ain't that the truth look at my join date.

Tom left the site in 2007 so I hope you were a reader before your join date.
 

Makaveli

Splendid
Ive been here since fall of 96 but the site has my join date way off.
Even tried the original Tom'sDelfi forums which were a joke.

ahh thanks for the update.

So there was some jank with the join dates once they moved the site from old to new.

i've been a reader since 99 but my join date is accurate.
 
Last edited:

Kamen Rider Blade

Distinguished
Dec 2, 2013
1,280
810
20,060
It isn't surprising that the 5950x won, I think they're pointing out the margin of victory. The 5950x finished 16.3% ahead of the 10900k with a 12.25% clock deficit. That would indicate a nearly 30% higher IPC than the 10900k, which would be higher than even AMD claimed using the 19% better than the 3000 series claim as a basis. This is for single threaded performance. In multi-threaded, the 5950x is obviously going to obliterate the 10900k.

It is best to "Under Promise and Over Deliver" than "Over Promise and Under Deliver".

I think AMD has learned from past mistakes in communications to the public.
 
It isn't surprising that the 5950x won, I think they're pointing out the margin of victory. The 5950x finished 16.3% ahead of the 10900k with a 12.25% clock deficit. That would indicate a nearly 30% higher IPC than the 10900k, which would be higher than even AMD claimed using the 19% better than the 3000 series claim as a basis. This is for single threaded performance. In multi-threaded, the 5950x is obviously going to obliterate the 10900k.
Don't forget that Zen 3's is 19% higher than Zen 2. Zen 2's IPC was already 7-10% higher than Sky Lake derivative CPUs.
 

Intel999

Honorable
Jan 13, 2015
17
2
10,525
Till actual benchmarks show up, these passmark results with 1 CPU of unknown settings should be ignored. Everything else you'll compare it to would have thousands of results averaged
Dude, you sound about as oblivious on here as you do when posting on Reddit.

There is very little unbelievable about this Passmark posting.

I know it's not easy to accept reality when everything your fanboyism has taught you to believe comes crashing down around you.

I suspect your views on GPUs leans toward the green team, as well.

May I suggest you pop a Xanax prior to Wednesday's launch event for the 6000 series of GPUs. You are going to need it. Lol
 
  • Like
Reactions: Jim90