Intel's upcoming Alder Lake-S lineup of processors is shaping up to be a rather good generational improvement. With wonders of the Intel 7 process, previously called 10 nm Enhanced SuperFin (10ESF), the processor lineup will deliver new hybrid technology, mixing new big and small cores into one package. Today, some new CPU-Z validation tests have shown up for the Intel Core i5-12600K CPU, which directly replaces the previous Core i5-11600K Rocket Lake model. With six high-performance Golden Cove and four efficient Gracemont cores, the Core i5-12600K CPU is a ten-core design with 16 threads. And compared to the 6C/12T i5-12600K CPU, the performance is much higher.
According to CPU-Z scores, the new Alder Lake processor scored 7220 and 7156 points for a multi-threaded benchmark in two tests. Compare this to the previous-generation model, which scores 4731 points, and the new chip is almost 50% faster. According to CPU-Z, the new CPU achieved this while running at a boost frequency of 4.5 GHz to 4.7 GHz.
58 Comments on Intel Core i5-12600K CPU-Z Scores Show 50% Higher Multi-Threaded Results Than i5-11600K
You already know this so you are just trolling.
browser.geekbench.com/v5/cpu/compare/9510991?baseline=9510991
For example, I took at top 10% scoring 5950X and compared it to the 12900K geekbench. From that, I can tell that the 5950X is better at text compression and certain physics like n-body and rigid body physics. 5950X is losing in things like AES, pdf rendering, html5, navigation, and ray tracing.
How that will affect real-world apps right now is anyone's guess, but I'll bet that the 12900K is going to be king of the web browser benchmarks. Good chance it will win on things like Blender and Cinebench as well. Because of those physics scores, It will probably lose on physics benchmarks like in Firestrike CPU and such, and may not be a lot faster in games as well. OFC, unless you're packing a 3090 and running at 1080p, you probably won't see any difference in games anyway.
I'll also say, vs my mildly tweaked and power unlocked 10850K build, the 12900K score posted beat my best scores in every single sub-category of single and multi-thread - in a few categories it was solidly twice as fast.
EDIT: oops, it’s actually fake quad. Well it’ll look great in the marketing material regardless.
www.techpowerup.com/272715/first-signs-of-amd-zen-3-vermeer-cpus-surface-ryzen-7-5800x-tested
+50%
= Less impressive than you imply.
I don't care either way, as long as what I've got continues to be useful (To me) I'm not gonna worry about it.
To me the big question is how are they doing with their GPU's? They have some new and old blood mixed into that division.. should be interesting what they can come up with.
I'm not a fanboy of either, but I do enjoy a good fight like any other red blooded male. I look forward to purchasing a new Intel, or AMD rig when the time is right..
- got rid of its entire testing department - WE are the testing department and are already reaping the rewards many times over
- helped keep Bulldozer down on its luck in Win 7 (yes, not great, but it certainly had a bit more potential in it)
- had questionable compatibility with Ryzen 1000 + Win 10 scheduler (AMD said it was just fine, but it was Ryan Shrout (lol), and half-truth damage control is not new from AMD)
- had godawful compatibility with Ryzen 3000 until Q1 2020+ on Win 10 scheduler (yes, half of it was on AMD, but just look in HWInfo for Windows' own core hierarchy)
- 'fixed' what ain't broke with Ryzen 5000 + Win 11, then didn't actually fix it (even after L3 and CPPC2 fixes, still benches side-by-side slower than 10, and L3 variance is still all over the map)
Aside from Ryzen 3000, closest we've got was Lakefield. No one cared so no one bothered to investigate Lakefield - just plain bad, or Win 10 scheduler strikes again? Performance was appalling, Win 10 scheduler deliberately kept the P-core out of action...........which is the exact same philosophy Intel confirms for Alder Lake, albeit with better cores...........I wanted some adventure, and you know what? for the most part it is just like running an Intel lol.. unfortunately not everyone is so lucky. And this whole 11 debacle is just sad. It just goes to show it doesn't matter how much money you have, or have much of the market you own.. you can still be a poor sport lol. Maybe one day I will be a Linux user.. but not today.
I was wrong.
They made it in the last minute.
Another day another Intel
PRLeakMS made the L3 cache fix as part of 282, and it failed miserably (despite non-ryzen owners and outlets reporting it 'fixed'). No, AIDA is no hero (far from it the pos software it is), but compared to performance on 10, the L3 variations on 11 are ridiculous. Run-to-run I still regularly see anywhere from 100-800GB/s and up to 15ns. On 10, it's a consistent 700-800 and 10.4ns every run, back to back, dozens and dozens of runs. I dual boot 21H1 and 282, and this "fixed" L3 perf is far worse than when the APUs were bugged on L3 earlier 2021.
And so the performance regressions persist. I still haven't been able to match Win 10 results on ST perf yet, and some games bizarrely clock much lower than 10 with no visible effects.
You are right, I'm not making excuses for AMD's pathetic firmware-writing and bugfixing abilities. But the way MS conducts themselves these days (remember their neverending struggle with Print Spooler?) affects hardware performance now, and the whole Ryzen 5000 + 11 debacle (where literally it works 100% fine on 10) doesn't bode well for the future. Part of me wants to believe that Intel's close relationship with MS should be able to avoid major issues, but Intel wasn't 100% ready with microcode for the RKL launch, either. Nevertheless the stakes are high for ADL since there is no other OS to fall back on, so fingers crossed both the scheduler and Thread Director hit the ground running.
I wonder if this is DDR4 or DDR5