Post your Cpumark99 scores

Page 18 - Seeking answers? Join the AnandTech community: where nearly half-a-million members share solutions and discuss the latest tech.

looncraz

Senior member
Sep 12, 2011
722
1,651
136
Well it appears we have topped out CPUmark99. The best any CPU will do is 5.60Mhz/CPUmark99. Meaning it requires 5.60 MHz from the CPU core to generate 1 CPUmark99. IPC gains with this benchmark basically flattened out after Haswell (5.8)

I did a test with a 5950X at 5GHz and scored 903... within 1.2% of the same 5.6Mhz/CPUMark99, so close enough to say you seem to be right. Is CPUMark99 x87 heavy?
 

Hulk

Diamond Member
Oct 9, 1999
4,191
1,975
136
Here is a chart of average CPUmark99 scores represented by the efficiency of this benchmark vs the core. The chart shows the MHz required for the core to generate 1 CPUmark99. Gains with modern cores are insignificant compared to real world gains so this benchmark is pretty much useless now. I've been recording scores for 21 years so I figured I'd present my data one last time. Core results are averages of scores I've recorded.
The benchmark is not multithreaded and obviously does not have much parallelism so modern, wider cores don't show much gain since around the introduction of Haswell.

1615068108640.png
 
Last edited:

AdamK47

Lifer
Oct 9, 1999
15,168
2,829
126
11900K @ 5.2GHz
nSOwFDI.png
 

Hulk

Diamond Member
Oct 9, 1999
4,191
1,975
136
11900K @ 5.2GHz
nSOwFDI.png

I assume that score is 1000, or is it a strange way of writing 1003?
Are you sure the clock wasn't 5.3 with Adaptive Boost Technology?
Would love to see the Handbrake score in our thread in this forum!

Nice!
 

AdamK47

Lifer
Oct 9, 1999
15,168
2,829
126
Yes, I forgot. This was 5.3GHz two cores and 5.2GHz on all core. This was most likely at 5.3GHz.

Once over 999 it represents 1000 as 1e+003.
 
Last edited:
  • Like
Reactions: Hulk

Hulk

Diamond Member
Oct 9, 1999
4,191
1,975
136
Yes, I forgot. This was 5.3GHz two cores and 5.2GHz on all core. This was most likely at 5.3GHz.

Once over 999 it represents 1000 as 1e+000, so this score is 1003.

Wow, very cool. You are like Chuck Yeager breaking the 1000 point "sound barrier" for this old benchmark!

For whatever reason Rocket Lake is very efficient at this one.
 

Miramonti

Lifer
Aug 26, 2000
28,651
100
91
11900K @ 5.2GHz
nSOwFDI.png
Never would have imagined it would take this long to get to 1k, but congrats, that's awesome. I'm still stuck on my single-threaded windows work app (64bit now), which was the reason for starting this thread, but looks like this app has run it's course.
 
Apr 20, 2008
10,161
984
126
5500u: 712 exactly three times, 714 when affinity is set to only one thread (CPU0). Picked up an open box Asus Zenbook 14 for $350. CPU and GPU both are better than I thought.

Screenshot 2022-09-12 163106.png
 

Hulk

Diamond Member
Oct 9, 1999
4,191
1,975
136
CPUmark99 has topped out on modern architecture. By "topped out" I mean no further IPC increases in this benchmark since Skylake.
In fact, I'm seeing a regression with Golden Cove and Gracemont with my initial testing.
 
  • Like
Reactions: Makaveli
Apr 20, 2008
10,161
984
126
6800U: 840

From an Asus Zenbook 13 S. Effective clock was between 4600~4650MHz during the run

What's the build quality like on the 13? I just picked up a clearanced/open box 14 and its close to being as premium as my M1 MacBook.
 

thigobr

Senior member
Sep 4, 2016
231
165
116
Build quality is very good, it's indeed very close to the Macbooks. I was torn between the M1/M2 Air and this 13 S OLED. I mainly got this because of Linux support. Battery life is great, very competitive with the Apple silicon. iGPU is also a big step from older Vega.
 

looncraz

Senior member
Sep 12, 2011
722
1,651
136
CPUMark99.png

Ryzen 9 7950X bone stock, but using a custom loop.

Running through WINE, though it doesn't appear to impact the score at all... Frequency was 5.88Ghz during most of the test - ran it three times, very little variation.