• Competitor rules

    Please remember that any mention of competitors, hinting at competitors or offering to provide details of competitors will result in an account suspension. The full rules can be found under the 'Terms and Rules' link in the bottom right corner of your screen. Just don't mention competitors in any way, shape or form and you'll be OK.

*** AMD "Zen 4" thread (inc AM5/APU discussion) ***

humbug said:
From your testing?

It has more per core L2 cache, that should amount to something.
----------

At this point, with all this now said and done, i still think Raptorlake will have an edge in performance, all be it a small one, both MT and ST.

AMD should still have an edge in efficiency, please don't bencher.

Gaming, no idea, Zen 4 also gets a doubling of L2, which should help, the clocks are very much higher than Zen 3, and there should be a boost from fast DDR5, put all that together and it looks interesting.

Once Zen 4X3d land tho, those will be 'thie' gaming chips.
Click to expand...
Yes I tested my 12900k at 5.8ghz ST, it scores similarly to the raptor lake leaks at various benchmarks (CPU z etc.). There is no IPC increase

I assume you are talking about cinebench and other heavy MT tasks, cause in most other workloads (gamging included) it never had the edge in efficiency.

Measuring efficiency at stock is useless in those tasks, whoever has lower limits wins. If Intel decides to ship their 13900k with a 50w power limit - boom - the most efficient CPU on the planet. That's silly though, it's just the wrong way of measuring efficiency. Who really cares about efficiency at stock power limits? Stupid metric
 
humbug said:
I take it then that the CPU-Monkey 13900K is complete nonsense?
Click to expand...
Are you talking about this? https://www.cpu-monkey.com/en/cpu-intel_core_i9_13900k

The numbers there don't make any sense. The 12900k at 5.5ghz gets 895 points in CPU Z, it has the 13900k at 893. But then in CBR23 it has it at 2267 compared to 2082 for the KS (which is at 5.5 ghz). I assume the numbers are taken with various clockspeeds / not a final product. All the leaks that appeared with a hwinfo screenshot (showing clockspeed) i measured a 0 to 1% IPC difference when I tested at same clocks.
 
Bencher said:
Are you talking about this? https://www.cpu-monkey.com/en/cpu-intel_core_i9_13900k

The numbers there don't make any sense. The 12900k at 5.5ghz gets 895 points in CPU Z, it has the 13900k at 893. But then in CBR23 it has it at 2267 compared to 2082 for the KS (which is at 5.5 ghz). I assume the numbers are taken with various clockspeeds / not a final product. All the leaks that appeared with a hwinfo screenshot (showing clockspeed) i measured a 0 to 1% IPC difference when I tested at same clocks.
Click to expand...

I was talking about this, not sure if its the same one.

Cinebench R20 (Single-Core) CPU benchmark list

Cinebench R20 (Single-Core) list with performance comparison of all CPUs in this benchmark
www.cpu-monkey.com www.cpu-monkey.com

12900K: 779
13900K: 888 (+14%)
 
humbug said:
AMD should still have an edge in efficiency
Click to expand...
Efficiency depends on the workload. The chiplet design is really great for high utilization workloads but really bad for low utilization. My 5950x was consuming 65-70w while watching a video using madVR and with some downloads in the background. I downgraded to a 5700G (monolithic design) as I didn't need the CPU power anymore, and that same workload consumes 20w.
 
humbug said:
Are you sure?

Stock....

Guru3D has it at 753
TweakTown has it at 754
Anand have it at 758
Click to expand...
Unless they measure the actual frequency during the run the numbers are worthless. Intel's ST boost is very finnicky, anything running on the background will drop the boost. When I test for ST I clock all cores to the frequency I wanna measure, or if that's too high (5.8ghz for example) I clock the 3 core frequency to the clockspeed i wanna test.

The numbers you are quoting are probably at 4.9-5.0 ghz speeds.
 
pretty sure they run them on test builds with clean windows and at STOCK , you can kind of get the idea if multiple reviews have around the same scores,
 
So, this seems to confirm the 13700K turbo boost speeds (5.4ghz):
www.techpowerup.com

Canadian Retailer Listings of "Raptor Lake" Confirm Max Boost Frequencies of Key SKUs

The flagship Core i9-13900K/KF "Raptor Lake" processor could come with a maximum boost frequency as high as 5.80 GHz, according to early store listings by a Canadian retailer that mentions the processor's retail SKU. This would be the highest possible clock speed sustained by the "Raptor Cove"...
www.techpowerup.com www.techpowerup.com

Based on that, I would only expect it to do a little better than the 12900KS (5.3ghz turbo boost) in single core benchmarks, perhaps about 810 in Cinebench R20 (about 2% higher than the 12900KS score).

At upto 5.8ghz, the 13900K single core score doesn't look too extreme, but might be clocked slightly higher than 5.8ghz. At these frequencies, I doubt it will be priced affordably.

Another possibility, is that the Cinebench R20 benchmark benefits from the higher L3 cache that the 13900K has, compared to all other Golden Cove CPUs.

At 5.1ghz turbo boost, it looks like the 13600K will be up against the 7700X, and should get a similar single core score to the 12900f.

It will be interesting to see if any AM5 motherboards can utilize any features to provide boosts to single Zen 4 cores (or even, higher frequencies across all cores), beyond the turbo boost speeds you get out of the box.

It seems entirely possible that a Zen 4 CPU clocked 300-400mhz higher than the 7700X could overtake the 13700K in single core performance...
 
12600K 5,1ghzP 4ghzE vs 7700X

bb0d6edaf651bc5da837b9df77ef1aa2e1fd7bb7824021606a56757e3f44f0fc.jpg
 
Seems that Y-Cruncher had been optimized for Zen 4 with AVX-512.

www.techpowerup.com

Latest Y-Cruncher Version Comes with "Zen 4" and AVX512 Optimization

Y-Cruncher is a multi-threaded Pi calculation benchmark. Its author, Alexander Yee, has access to an AMD Ryzen 9 7950X 16-core/32-thread sample, and has developed the latest version 0.7.10 of the Y-Cruncher binary with optimization for the "Zen 4" microarchitecture, and to take advantage of the...
www.techpowerup.com www.techpowerup.com

Will still be the go to for stability testing, and nice to see the developer getting it ready pre-launch.
 
You must log in or register to reply here.
Back
Top Bottom