• Welcome to Overclockers Forums! Join us to reply in threads, receive reduced ads, and to customize your site experience!

New Cinebench 2024

Overclockers is supported by our readers. When you click a link to make a purchase, we may earn a commission. Learn More.
Was about to post the same. Didn't spot this thread earlier as didn't think to look in the AMD section!

Some technical notes - focusing on PC. From https://www.maxon.net/en/tech-info-cinebench

Minimum Requirements (Windows)

  • 16 GB of RAM
  • 64-bit Intel or AMD cpu with AVX2 support or Windows 11 on Snapdragon compute platform or ARM v8.1 64-bit CPU
  • NVIDIA GPU with CUDA compute capability 5.0 or higher and 8 GB VRAM, or
  • AMD "Navi" or "Vega" GPU or later with HIP capability and 8 GB VRAM or more (see GPU list below)
On AMD GPU support, it looks like they only tested it on some models from RDNA1 onwards, but expect Vega dGPUs to work. I'd love to see someone try this on a Vega APU.

  • Compared to Cinebench R23 the computational effort for the scene increased sixfold in the multithreaded rendering test. This reflects the CPU performance improvements as well as the increased demands artist have to deal with these days.

  • The higher complexity of the scene also means the memory footprint has increased threefold. Depending on your CPU configuration Cinebench 2024 will need at least 6.5 - 8.5 GB RAM. The minimum memory requirement is therefore set to 16 GB (while macOS can execute the CPU test also on 8 GB machines – with significant influence of paging – Windows usually has several GB assigned to non-purgeable memory which prevents the execution of Cinebench 2024 on 8 GB machines).
While they say 16GB is required, I'd guess if you have an odd ball system with 12GB it might work too.

I've yet to start my own testing with it. I'm seeing early reports from elsewhere that this might be the first version to be impacted by ram performance.
 
Right?! No pics of results?!!! :p

Cool... will add this to new mobo reviews in October.....

I've yet to start my own testing with it. I'm seeing early reports from elsewhere that this might be the first version to be impacted by ram performance.
For testing purposes, RAM is kept at the same speed, so not terribly relevant within that scope. Though it would make comparisons between reviews more difficult for sanity checks, lol.
Some technical notes - focusing on PC.
This runs on something other than a PC?




Will move this to the proper section........


EDIT: Here's a DL link as well (1.3GB FYI).............................. https://www.maxon.net/en/downloads/cinebench-2024-downloads
 
Last edited:
Right?! No pics of results?!!! :p
My first run on 7920X. I had task manager up and towards the end I saw "System" take around 4% CPU so I don't consider this a clean run. Will repeat later when I have more time. A run takes a while now.

cb2024-1r.png

For testing purposes, RAM is kept at the same speed, so not terribly relevant within that scope. Though it would make comparisons between reviews more difficult for sanity checks, lol.
I've considered Cinebench up to R23 as a joke of a benchmark since when runs are properly optimised, scores can be predicted quite well from just knowing cores, clocks and CPU architecture used. Now that the workload may be getting big enough to start impacting ram performance makes it more interesting.

This runs on something other than a PC?
Mac people can join in too, as well as Windows on Arm.
 
Mac people can join in too, as well as Windows on Arm.
Gotcha! :)

That's all a PC to me! Just like a ICE car, RX7 (ICE but a Wankel), and Tesla (electric) are all cars, even if they work a bit differently. :rofl: :chair: :escape:

I've considered Cinebench up to R23 as a joke of a benchmark since when runs are properly optimised, scores can be predicted quite well from just knowing cores, clocks and CPU architecture used.
I'd call that consistent and good scaling in its runs. :p


and it don't show all the boxes as it renders
....running the CPU multi-core, I see the squares/activity as it renders. However, there are only two as opposed to the number of threads on the CPU as past versions ran.


Here's a screeny of a 13900K running CPU multi-core (stock with 3x120mm AIO) with DDR5-7200 CL36). Dirty run too, with FFox and 13 tabs up, lol. It took longer to extract/uncompress (to 2.3GB) than to DL, lol. It does take forever to run (~10+ mins on my config), wow........not even going to try single core now.

That's going to extend my benchmarking run time by quite a bit, though. Something that took around a minute (x3) is now going to take 30 mins or so for three runs.



1693999065875.png

PS - In a non-apples to apples comparison, my other 13900K system with DDR5-6000 CL30 RAM (unlocked to 253W PL1/2 - as is the run from above) scored 1949.... but that was also a clean run with idle OS/no apps up.

PPS - Single core on 13900K was 128.
 
Last edited:
That's all a PC to me!
I think to many, saying PC implies Windows/x86. While the others are personal computers in the wider sense, they're not commonly called PCs.

That's going to extend my benchmarking run time by quite a bit, though. Something that took around a minute (x3) is now going to take 30 mins or so for three runs.
If you go into advanced menu you can turn off the 10 minute requirement. If the cooling is good, then a cold run without warmup should be consistent. It still takes a while though. I've not timed it.
 
I think to many, saying PC implies Windows/x86. While the others are personal computers in the wider sense, they're not commonly called PCs.
Depends on who you ask! :p

If you go into advanced menu you can turn off the 10 minute requirement. If the cooling is good, then a cold run without warmup should be consistent. It still takes a while though. I've not timed it.
Good to know. I don't recall needing to do that previously. Annoying.
 
Logitech GHUB, PCPanel Software and Stream Deck scftware running in the background, otherwise everything else is closed.

I don't know about why my 4090 score is so low when it seems other 4090s out there are hitting 37k - 39k. Maybe I need to update to the latest driver....

Cinebench2024.jpg
 
@Celeron_Phreak - I got 35.1k with a 4090 Suprim using the latest driver, FYI.

Where are you seeing people hit 37k+?

EDIT: Your 13,900k score is way higher than mine though.
 
Last edited:
I see one person hit that score (end of the thread)... it's 4 pages long and half the people don't post the hardware and/or their sig is wrong, so more is hard to find, lol.
 
I see one person hit that score (end of the thread)... it's 4 pages long and half the people don't post the hardware and/or their sig is wrong, so more is hard to find, lol.
Whoops, I was bleary eyed when I went over what I missed this am :)
 
Still... that's A LOT higher than 35.1k...

Not sure what's up with mine or why he's so much higher (mine is watercooled too and tends to sustain clocks better than air).
 
@Celeron_Phreak - I got 35.1k with a 4090 Suprim using the latest driver, FYI.

Where are you seeing people hit 37k+?

EDIT: Your 13,900k score is way higher than mine though.
I've seen a couple of them on reddit, and a couple of youtube videos. It looks like 10 minute runs, but I assume they have some overclocks set in place. A lot of these high scores of 37k+ are strix cards. It looks like MSI, zotac and PNY cards hit the 35-36k mark, stock, pretty regularly.

I find it crazy that with what little OC can be achieved from today's cards, a 4090 can hit 39 or even 40k on this test.

I'm not sure why my 13900k score is so high, honestly. I've been seeing the normal 13900k score is around 2200. I don't have any OC applied and this windows install is at least 9 months old (daily driver), and I didn't even do a reboot.

I am now very interested in seeing what difference the Z790 Dark K|ngp|n board will do when I swap it out with a fresh SSD.
 
I don't remember the Z690's default behavior, but I'd guess it's a full 253/253 on PL1/2. Not sure what mine is... "Asus enhanced" or something on a ROG Extreme...

.... really want to swap it out with the Z790 Dark Kingpin, but need faster ports. :(
 
Back