Question Geekbench 6 released and calibrated against Core i7-12700

[igor_kavinski]

Geekbench 6 - Geekbench Blog

www.geekbench.com

Weird choice of baseline CPU and even weird is that the baseline score is 2500.

i7-12700 does hardly 2000 in GB5 with the fastest DDR5.

EDIT: Gideon's browser extension/add-on: https://forums.anandtech.com/thread...d-against-core-i7-12700.2610597/post-41452115

Helps you make sense of the results if you are left scratching your head about why a certain test is fast on a certain CPU.

Super useful!

Sample screenshot:

 

[Markfw]

255H vs. 7840U

LENOVO 83KF vs HP HP Pavilion Plus Laptop 14-ey0xxx - Geekbench

(16 cores @ ??W vs 8 cores @ ~40W)

How about an actual review instead of just geekbench 6 ? I think that would be more useful.

 

[gdansk]

How about an actual review instead of just geekbench 6 ? I think that would be more useful.

I do not have any Arrow Lake H laptop and it's the Geekbench 6 thread.

 

[Markfw]

I do not have any Arrow Lake H laptop and it's the Geekbench 6 thread.

I realize that. But those 2 laptops don't seem comparable, so doing your own is fine, but doing it vs a random laptop seems meaningless.

 

[gdansk]

I realize that. But those 2 laptops don't seem comparable, so doing your own is fine, but doing it vs a random laptop seems meaningless.

That's the point. It isn't comparable and yet within a few percent in GB6 1T. A bit of a joke.

 

[igor_kavinski]

285H vs. Strix Halo: https://browser.geekbench.com/v6/cpu/compare/9623017?baseline=9460276

Winning in ST but getting clobbered by AMD in all MT tests except two.

Lack of HT is hurting. Good job, Intel! /s

 

[igor_kavinski]

285H vs. 185H: https://browser.geekbench.com/v6/cpu/compare/9623017?baseline=9655356

 

[igor_kavinski]

Picked the two most performant recent Windows scores from the competing x86 camps: https://browser.geekbench.com/v6/cpu/compare/9843770?baseline=10002777

Geekbench needs an additional test to mimic gaming performance coz I don't see anything Arrow Lake is really terrible at from those subtests. If Geekbench is showing the widely panned Arrow Lake in such good light, then it's safe to say that Geekbench may also be failing to highlight weaknesses in M4.

Comparison with M4: https://browser.geekbench.com/v6/cpu/compare/9820610?baseline=9843770

Interestingly, GB 6.4 could be nearing release: https://browser.geekbench.com/v6/cpu/9820610

 

[Det0x]

Picked the two most performant recent Windows scores from the competing x86 camps: https://browser.geekbench.com/v6/cpu/compare/9843770?baseline=10002777

Geekbench needs an additional test to mimic gaming performance coz I don't see anything Arrow Lake is really terrible at from those subtests. If Geekbench is showing the widely panned Arrow Lake in such good light, then it's safe to say that Geekbench may also be failing to highlight weaknesses in M4.

Comparison with M4: https://browser.geekbench.com/v6/cpu/compare/9820610?baseline=9843770

Interestingly, GB 6.4 could be nearing release: https://browser.geekbench.com/v6/cpu/9820610

Why use that junk 9950X score vs known guy from overclock.net? 🤣

 

[igor_kavinski]

Why use that junk 9950X score vs known guy from overclock.net? 🤣

OK, I compared with your score: https://browser.geekbench.com/v6/cpu/compare/9843770?baseline=9430627

10 wins in MT for 285K.

Point is, GB6.3 isn't telling how bad 285K is in games. It's like Intel made sure to accelerate codepaths they knew would be used in CB R24 and Geekbench.

 

[Det0x]

OK, I compared with your score: https://browser.geekbench.com/v6/cpu/compare/9843770?baseline=9430627

10 wins in MT for 285K.

Point is, GB6.3 isn't telling how bad 285K is in games. It's like Intel made sure to accelerate codepaths they knew would be used in CB R24 and Geekbench.

Yeah i kinda see your point and i agree
Looking at it from the other side, non of the benchmarks on hwbot shows how good X3D v-cache really is in gaming

Overclocking, overclocking, and much more! Like overclocking.

HWBOT is a site dedicated to overclocking. We promote overclocking achievements and competitions for professionals as well as enthusiasts with rankings and a huge hardware database.

hwbot.org

Guess we need some new benchmarks that is a better metric for "real gaming performance".. Not even 3dmark is a good comparison for how games really behave
(3dmark cpu score is acting like a mini render benchmark where more cores = more better)

 

[igor_kavinski]

To Be Filled By O.E.M. ROMED8-2T vs ASUS System Product Name - Geekbench

browser.geekbench.com

Is the 7373X faster purely due to the V-cache? Please dissect and discuss.

 

[MS_AT]

To Be Filled By O.E.M. ROMED8-2T vs ASUS System Product Name - Geekbench

browser.geekbench.com

Is the 7373X faster purely due to the V-cache? Please dissect and discuss.

Different geekbench version. Epyc will have more DRAM channels. If you would like to single out cache effects, compare 9800x3d overclocked to same clock as 9700x for the same geekbench version and OS. That should be the most apple to apple comparison I guess.

 

[poke01]

Geekbench 6.4 - Geekbench Blog

www.geekbench.com

 

[poke01]

Geekbench 6.4 - Geekbench Blog

www.geekbench.com

Very good improvements. @igor_kavinski have a look and test

 

[igor_kavinski]

Very good improvements. @igor_kavinski have a look and test

The only one I can try to test is the AVX-512 detection. Let me try to do get it done in a few hours if possible. Don't got no M4 or RISC-V CPU to test their other functionality.

 

[poke01]

The only one I can try to test is the AVX-512 detection. Let me try to do get it done in a few hours if possible. Don't got no M4 or RISC-V CPU to test their other functionality.

all good, I'll post my 7700X result too.

@Eug , can you please test your M4 mini?

 

[poke01]

Gigabyte Technology Co., Ltd. B650M GAMING X AX - Geekbench

Benchmark results for a Gigabyte Technology Co., Ltd. B650M GAMING X AX with an AMD Ryzen 7 7700X processor.

browser.geekbench.com

these are the instructions Zen4 uses in GB6.4

 

[poke01]

M4 Pro. Got this result from a random run

 

[igor_kavinski]

Geekbench 6.4 SME support now requires CPUs to either implement SME instructions or implement SME and SME2 instructions, improving compatibility and performance for upcoming CPU releases.

This could only mean one thing. Mediatek and Nvidia's SoC launch is imminent.

 

[Eug]

@Eug , can you please test your M4 mini?

24 GB M4 Mac mini (10-core - 4P/6E), macOS 15.3 - Geekbench 6.4.

Mac mini (2024) - Geekbench

Benchmark results for a Mac mini (2024) with an Apple M4 processor.

browser.geekbench.com

Mac mini (2024) - Geekbench

Benchmark results for a Mac mini (2024) with an Apple M4 processor.

browser.geekbench.com

8 GB M4 iPad Pro (9-core - 3P/6E), iPadOS 18.3 - Geekbench 6.3, since 6.4 unavailable.

 

[igor_kavinski]

GB 6.4 vs. 6.3 on Ice Lake Xeon: https://browser.geekbench.com/v6/cpu/compare/10185437?baseline=10184875





With 7-Max: https://browser.geekbench.com/v6/cpu/compare/10185557?baseline=10185095





Maybe Intel crept in some compiler optimizations?

 

[Nothingness]

Maybe Intel crept in some compiler optimizations?

To clarify : Geekbench doesn't use Intel compiler as far as I know (IIRC it's clang for all platforms).

 

[igor_kavinski]

To clarify : Geekbench doesn't use Intel compiler as far as I know (IIRC it's clang for all platforms).

Maybe Intel engineers submitted optimization patches to the Clang codebase?

 

[Doug S]

Gigabyte Technology Co., Ltd. B650M GAMING X AX - Geekbench

Benchmark results for a Gigabyte Technology Co., Ltd. B650M GAMING X AX with an AMD Ryzen 7 7700X processor.

browser.geekbench.com

View attachment 115830
these are the instructions Zen4 uses in GB6.4

I really wish the main results page listed those somewhere, or had a link you could click to get to that. I have no idea using your first link to the Geekbench results how to get to the information you quoted after showing the "value" list.

 

[igor_kavinski]

I have no idea using your first link to the Geekbench results how to get to the information you quoted after showing the "value" list.

Need to be logged in for that to show: https://browser.geekbench.com/v6/cpu/10179944.gb6

Append ".gb6" to any GB6 result.



Sadly, still doesn't show which subtest is using which ISA extension(s).