AMD Ryzen 5 9600X ES Overclocked to 5.70 GHz All-core

[btarunr]

An AMD Ryzen 5 9600X "Zen 5" processor engineering sample has been overclocked to 5.70 GHz on all cores, and put through the CPU-Z Benchmark. Here, the chip is found to score 871 points in the single-thread benchmark, and 7096 points in the multithreaded benchmark. The overclock to 5.70 GHz is significant, as this is the maximum boost frequency of the upcoming Ryzen 9 9950X flagship part. The single-thread benchmark highlights that "Zen 5" has a similar IPC to the "Raptor Cove" P-core of the Intel Core i9-14900K processor, while the 7096 points multithreaded score is higher than the that of the Ryzen 7 5800X "Zen 3," meaning that AMD is overcoming the deficit of two whole cores (33% of the core-count of the 9600X) with just IPC and faster memory. AMD is expected to launch the Ryzen 9000 series desktop processors in July 2024.



View at TechPowerUp Main Site | Source

 

[Prima.Vera]

OK. What is the score under normal operation?

 

[Bwaze]

"871 points in the single-thread benchmark"

That's lower than stock 13900K. I know that all core overclocking of Zen processors significantly lowers their single thread results, but come on, this is supposedly max boost frequency...

 

[Jism]

"871 points in the single-thread benchmark"

That's lower than stock 13900K. I know that all core overclocking of Zen processors significantly lowers their single thread results, but come on, this is supposedly max boost frequency...

CPU-Z is using a outdated benchmark. It scores well on intel but does not take advantage of (new) codepath on ZEN CPU's.

 

[Fouquin]

"871 points in the single-thread benchmark"

That's lower than stock 13900K. I know that all core overclocking of Zen processors significantly lowers their single thread results, but come on, this is supposedly max boost frequency...

A 3.5% difference in a benchmark that swings either way seemingly at random (See: Zen 1's CPU-Z results) despite what the effective IPC actually is. Curious about the power draw though, CPU-Z's 17.01 bench really does nothing to push Ryzen chips that hard, usually 10-15% below TDP even on, "stress test".

 

[londiste]

What's the power consumption like at that frequency?

 

[Crackong]

Don't worry.
CPU-Z fixed this bug at Zen1 era

 

[usiname]

"871 points in the single-thread benchmark"

That's lower than stock 13900K. I know that all core overclocking of Zen processors significantly lowers their single thread results, but come on, this is supposedly max boost frequency...

CPU-Z's benchmark is just userbenchmark 2, with countless updates that make the ryzen to perform worse since gen 1. In the old version from 2016 before the first ryzen the Zen 4 is faster in single core than 13th and 14th gen Intel

 

[JWNoctis]

Meanwhile, Apple's M3/M4 struts about like a heron among a flock of geese, only to be hampered by their closed ecosystem and so-far limited market.

Both AMD and Intel need to raise the level of their game. More clock and IPC certainly would be a step in the right direction.

 

[londiste]

Meanwhile, Apple's M3/M4 struts about like a heron among a flock of geese, only to be hampered by their closed ecosystem and so-far limited market.
Both AMD and Intel need to raise the level of their game. More clock and IPC certainly would be a step in the right direction.

Where are Apple M3 in terms of clock speeds? 4GHz? Especially AMD but also Intel chips are also pretty damn efficient at that frequency

 

[ratirt]

Where are Apple M3 in terms of clock speeds? 4GHz? Especially AMD but also Intel chips are also pretty damn efficient at that frequency

No to mention, 4 efficiency cores 2.4 GHZ max and 4 performance cores 4.06Ghz max for an 8core chip, and people are amazed about the power consumption.
Intel would probably smash it with 4p+4e at that frequency running on an AAA batteries pff


Still don't understand, why would the OC Ryzen? Let the board and CPU manage the speed and power. That is the best way for Ryzen's in my opinion. Other you gain here and lose there. Pretty sure it is not worth it.

 

[Metroid]

My ryzen 5900x at 3.9ghz v0.9 does st = 534, mt = 7900. It has 12 cores 24 threads, meaning the new 9600x with 6 cores 12 threads is almost on par with my 5900x, pretty good. I wonder clock per clock how would that do.

 

[R0H1T]

An 8c/16t zen 5/6(c) only "light" cores with say LPDDR6 should be interesting wrt efficiency. People don't realize how controlling the entire ecosystem by Apple is such a massive advantage for them!

 

[usiname]

Still don't understand, why would the OC Ryzen? Let the board and CPU manage the speed and power. That is the best way for Ryzen's in my opinion. Other you gain here and lose there. Pretty sure it is not worth it.

"Let the board and CPU manage the speed and power." This is the worst possible, manual OC or undervolt is way better in terms of performance and power consumption
Example with my 7600x
Default 1.35V, 5.2Ghz max single core 5.1Ghz all core, 90W
Fixed voltage 1.2V, 5.4Ghz single and all core, 77W
Fixed voltage 1V, 5Ghz single and all core, 43W
This is tested in cinebench

 

[R0H1T]

The light cores are probably the most energy efficient in AMD's lineup right now!

 

[ratirt]

"Let the board and CPU manage the speed and power." This is the worst possible, manual OC or undervolt is way better in terms of performance and power consumption

With my 5800x OC is pointless. Undervolting can be beneficial but it does not work in a way to make a noticeable difference and can't be applied to all CPUs with the same values. That is not the case. Leaving the defaults gives you balanced performance and I'm sure this is the best way to go.
I really can't see why this particular situation would change to better for upcoming Ryzens.

 

[Gigaherz]

I know that all core overclocking of Zen processors significantly lowers their single thread results,

It doesnt. You just most of the time end up on a lower allcore frequency than sinclecore cause of stability - doesnt automatically mean allcore oc lowers SC performance tho.
If you get static allcore oc above stock boost on allcore, you will ofc achieve higher SC results too.

 

[usiname]

With my 5800x OC is pointless. Undervolting can be beneficial but it does not work in a way to make a noticeable difference and can't be applied to all CPUs with the same values. That is not the case. Leaving the defaults gives you balanced performance and I'm sure this is the best way to go.
I really can't see why this particular situation would change to better for upcoming Ryzens.

I had ryzen 5500 and the situation was the same as with Zen 4. I could choose between 10% lower powerconsumption with 100-200 Mhz over the default or ~30-40% lower power consumption for 5% lower performance. But yes, Zen4 is better in this than Zen3

 

[R0H1T]

If you get static allcore oc above stock boost on allcore, you will ofc achieve higher SC results too.

You generally get higher clocks or ST scores with PBO2+offset it's also the most efficient way to OC as far as I'm aware. Most reviews I've seen always put PBO scores at or near the top as compared to manual OC, so no unless you're setting just the one core to boost ridiculously high & the rest like -500MHz compared to that manual OC will not net you the best SC results.

 

[Dr. Dro]

CPU-Z is using a outdated benchmark. It scores well on intel but does not take advantage of (new) codepath on ZEN CPU's.

This is not a codepath issue, it's that the 2017 benchmark (default on CPU-Z) is not vectorized. The 2019 ones should post great scores with a manual setting since they use AVX, as long as the chip can do it without thermal or current throttling. This is plenty fast for SSE, let's be realistic here.

CPU-Z's benchmark is just userbenchmark 2, with countless updates that make the ryzen to perform worse since gen 1. In the old version from 2016 before the first ryzen the Zen 4 is faster in single core than 13th and 14th gen Intel

Hold steady there fella. A little early for you AMD boys to be huffing the copium. There haven't been "countless updates to make Ryzen perform worse", there is no grand conspiracy at play here. This is just one benchmark out of many, and as others have pointed out in this thread, it's not universally representative of performance. I think this score is actually amazing for a Ryzen 5-class processor.

 

[usiname]

Hold steady there fella. A little early for you AMD boys to be huffing the copium. There haven't been "countless updates to make Ryzen perform worse", there is no grand conspiracy at play here. This is just one benchmark out of many, and as others have pointed out in this thread, it's not universally representative of performance. I think this score is actually amazing for a Ryzen 5-class processor.

Name one benchmark updated to favor AMD and I will name 5 updated to favor intel

 

[Carillon]

with just IPC and faster memory

and substantially higher clocks

 

[R0H1T]

There haven't been "countless updates to make Ryzen perform worse", there is no grand conspiracy at play here.

Just because you're paranoid doesn't mean

Spoiler

 

[marios15]

Just because you're paranoid doesn't mean

Spoiler

He's right though, CPUZ default benchmark algorithm was replaced with a different codepath to reduce scores on Ryzen (supposedly to match "expected performance") because the benchmark would run extremely fast on Zen microarchitecture due to its cache.

Imagine a game releasing a patch that made the game slower because it's "running too fast" on a new architecture

 

[Dr. Dro]

He's right though, CPUZ default benchmark algorithm was replaced with a different codepath to reduce scores on Ryzen (supposedly to match "expected performance") because the benchmark would run extremely fast on Zen microarchitecture due to its cache.

Imagine a game releasing a patch that made the game slower because it's "running too fast" on a new architecture

Should be simple enough to solve that dispute. Here's my score on the CPU-Z 2015 benchmark. It doesn't have the faintest clue about how Raptor Lake works, so I suspect these scores are about as worthless as it comes. Still of academic interest, considering the cache sizes on Raptor Lake are larger than even that of the first Ryzen Threadripper processors. Anyone on Zen 4 X3D to step up?

Download link: https://www.cpuid.com/downloads/cpu-z/cpu-z_1.78-en.zip