This is the 4.4 Ghz 4001 point run:
https://browser.geekbench.com/v6/cpu/6107586
Here is another gb5 result
https://browser.geekbench.com/v5/cpu/22466510
Geekbench 4 used to have a memory-latency test even (that gave some points):
https://browser.geekbench.com/v4/cpu/16671577
but the removed it in v5. I Guess they still measure it but it isn't visible in their json dump. Here is the output of a random gb6 run...
Whichever way you slice it, 20% better perf per core is a "run of the mill" incremental update from AMD. It's nowhere near "Osborning everything before it" or if it were, Zen4 with its 30% ST perf uplift would have been a bigger deal
Easy now, from the same results:
7845HX gets 356.92 - this is a chiplet-based 12 core (repurposed desktop) CPU (1 result in database)
7700X (desktop) gets 269.02 (24 results)
7940HS gets 224.22 (16 results)
7840U gets 216.09 (1 result) and 153.31 (for 5 results) not sure why they are separate...
The most interesting part for me is the 4.4Ghz boost in fridge /under liquid nitrogen (just check the .gb6 info)
If that's accurate and achieved just by lowering temps, macbook pros should be capable of reaching close to that on active cooling alone.
This should make M4pro very competitive...
Chips and Cheese looks at Crestmont (Meteor Lake's E-cores):
https://chipsandcheese.com/2024/05/13/meteor-lakes-e-cores-crestmont-makes-incremental-progress/
Right, I totally forgot that Prescott was already 64bit.
Tended to avoid desktop Intel CPUs ever since the P4 till Core 2. And as Core 1 (Yonah) was still 32 bit (and I remember the AMD64 saga very well) I somehow had a brainfart
Sorta OT bu interesting stuff is happening at Chips & Cheese:
https://chipsandcheese.com/2024/05/09/chips-and-cheese-state-of-the-union/
Among other things:
So it seems at least more high-quality tech-content is in order :)
BTW This surprised me:
NIce! I particularily hope the highlighted part is true. The 7600 series was quite a disappointing and having another generation of the same CU count, bandwidth and memory that the RX 6600 would be really sad.
I was initially afraid the 8xxx series would be handled by N44
I'd add the Intel thread, where there seems to be a consensus (based on a blurred picture) that the next big core adds 50-100% more resources to every backend feature, has gazillion-wide decode capability (with ucache), yet only delivers <=20% IPC uplift
Well hello CAMM2:
Coming to the Lenovo P1 gen 7 near you:
https://news.lenovo.com/pressroom/press-releases/lenovo-unveils-its-new-ai-ready-thinkpad-p1-gen-7-mobile-workstation/
I get what you are saying, but I actually faced this issue (when the naming scheme was relatively new) as laptops with the 7520U were available right next to a near-identical last-gen laptops with the 5625U, priced very closely.
The "higher is better" number immediately caught my eye (and got...
I know this has already been discussed to death, but i wanted to add a bit of reasoning what irks me about the naming-scheme, and how how this could be improved upon, even when model year is kept as the first number.
The main issue I have has to do how humans tend to read jumbled text (like...
Yeah, there are still plenty of ways to better hide the "OEM tax" number (that keeps rising ever year) The generation should be the first number or brought out totally separately e.g. "Gen 5"
It only works as the 8xxx series has no zen 3 parts. If it had a 8630U would be significantly worse...
Damn, I try not to let the hype get to me, but it still feels like the AMD hypetrain has yet-again set up quite insane expectations.
I feel like if Strix Point is released and only offers 15% ST gain, 15% GPU gain and 10% more battery life, people will be massively dissapointed.
That's correct. "Eco Mode" probably is covered/validated by mobo manufacturers and AMD. Undervolting most certainly isn't.
I'm not 100% sure about even manually setting the 3 values. (as there.is the same "Accept Potential Damage" dialog). It should be fine, as you'll be setting them stricter...
Sorry for editing the post, last minute. I'll condense the relevant parts out again:
So the 45W starting-point settings should be around:
PPT - 61 000 mW (this is fixed as the number is straight 1.35x "the desired TDP on the box". This is also the actual wattage you'll see in HWinfo at 100%...
"Eco mode" is nothing special. It's just a convenience "one click" button for setting 3 PBO values. You should still be able to set the same settings manually via PBO (pictures in the article):
https://overclock3d.net/reviews/cpu_mainboard/amd_eco_mode_guide_for_am5/
PBO has 3 settings:
PPT...
A side-slide to Machine Learning again. It looks that newer models are more-and-more "gamed" to fit the benchmarks (That will inevitably end in the training data, as it more-or-less uses the entire internet):
Phi3 and Mistral were the biggest culprits (losers) when run on a similar (but...
It would be wonderful to see a a cat-core derivate solution!
Yeah i was just thinking in the other thread, could it be connected to this patent:
https://www.techpowerup.com/283325/amd-files-patent-for-its-own-x86-hybrid-big-little-processor
Hopefully something on the igpu side as well, that...
WTH kind of promo video is this?
I though well of Hardware Canucks and I do get the need to do promo-videos in order to survive financially, but the timestamped part is where i lost it:
"Additionally with the Innovative dual boost technology 2 of those high-performance cores can surge to an...
Some stats about Strix, nothing earth-shatteringly new:
https://videocardz.com/newz/amd-ryzen-9050-strix-halo-specs-leaked-16-zen5-cores-and-40-rdna3-5-cus-lp5x-8000-memory-and-32mb-mall-cache
One interesting dibit is that Strix Halo still has full 32MB of L3 per CCD, so these look more and...
This result also reports more memory and seems to detect the cache a bit better (for all cores). The reported 5.1 Ghz base clock seems interesting, but is probably BS.
I don't get where you got that 21-25% IPC gain? Using linux a good 7840U would score a little over 2000 points at 5.0 Ghz ...
Well, the geekbench browser is full of hvm results that all seem to detect frequencies correctly, so it should.
Doesn't mean it did it for an early engineering sample though
Yeah I actually wanted to discuss it, as it was released yesterday but i said i'll end the OT.
Phi3's smaller versions are probably something they will embed to windows, but tune for os assistance rather than chat.
The 14b medium model truly looks excellent, but i'd wait for more benchmarks...
It's a spoffed score (probably messing with the clock), if you look at the frequency info it's "apparently" running at 11Ghz :p
https://browser.geekbench.com/v5/cpu/22423657.gb5 (Need to be logged in to see this)
Yeah, they are fine for many, but plenty of companies has strict-enough rules/contracts that disallow them tot share any company's (or client's) data with these platforms.
Agreed, that's why I mentioned they have some uses.
I'll end the OT talk now.
Business also run AI on local servers.
With ollama it's dead easy and now with llama 3 released, it's very-very competitive to commercial competitors. The biggest downside of it is, that the 70B parameter model requires at least 40GB of VRAM so you can't run it on any gaming GPUs (the 8B model...
OT, but Does Strix bring any relevant light load (e.g web browsing) efficiency improvements? Phoenix was otherwise awesome but rather dissapointing in that regard
I agree overall, but as an enthusiast, It's my hobby to moan about marketing decisions "nuking" engineering achievements. I'd much rather have 8 (or even 6) cores wit a still a single-core boost (leave 2-core for Elite if needed), than 10 cores and no boost.
And as adroc mentioned 7640U...
Finally a Notebookcheck review with a decent gen-on-gen battery life improvement (though the last gen had terrible scores):
https://www.notebookcheck.net/Core-Ultra-7-165U-performance-debut-Dell-Latitude-9450-2-in-1-review.825254.0.html#toc-6
Even a slight win over average 7840U laptops with...
This site uses cookies to help personalise content, tailor your experience and to keep you logged in if you register.
By continuing to use this site, you are consenting to our use of cookies.