4500 Geek6 ST and 220 Cine 24 ST? How do you think QCOM will be able to accomplish this jump in a single generation? Not even Z5 is rumored to reach these levels.
4500 Geek6 ST and 220 Cine 24 ST? How do you think QCOM will be able to accomplish this jump in a single generation? Not even Z5 is rumored to reach these levels.
Well, for one- I quadrupled the L2 cache from 12 MB to 48 MB. They could massively widen the core to achieve this. Of course this is all my speculation.4500 Geek6 ST and 220 Cine 24 ST? How do you think QCOM will be able to accomplish this jump in a single generation? Not even Z5 is rumored to reach these levels.
moar L2 doesn't help magically.Well, for one- I quadrupled the L2 cache from 12 MB to 48 MB. They could massively widen the core to achieve this. Of course this is all my speculation.
AMD is a very special case, their rules do not apply to anyone else.AMD's Zen5 core is said to bring a 40% uplift in SPEC INT RATE 2017. Since SPEC tends to correlate well with Geekbench, we are looking at a 40% uplift in Geekbench too. Phoenix Point -> Strix Point, is 2600 -> 3650.
But Strix Point is coming in H2 2024. Strix's successor with Zen6 is rumoured to be released sometime in 2026H1.
Considering previous Zen generations, Zen 6 will bring atleast a 15% performance uplift. So we are looking at about 4150 points in Geekbench 6, for Strix Point successor.
And this all that was for AMD's HS parts. As we know AMD also sells the HX parts, which tend to have 10-15% higher ST than the HS parts.
Lol. OkAMD is a very special case, their rules do not apply to anyone else.
Yeah, making a better CPU is a very comprehensive process that involves more than just piling up on cache.Lol. Ok
Yes disabling the boost clock was disgusting.The lack of boost except for a select few of high premium chips is really funny. QC may well carve out a WoA niche in offices where laptops are little more than glorified "thin" clients (and Surface gonna Surface anyway). But anybody calling that a serious attack on the current x86 domination of the traditional PC market has to be kidding.
Sure. They are watching this thread closely and besides, they will soon appoint you as their Chief Experience Officer so hopefully you will still engage in discussions here, unlike some of their architects subscribed to this thread who seldom say anything.Maybe Qualcomm intentionally leaked that in advance to gauge our reaction. Considering our reaction is negative, they will enable boost clock.
Doing just that and nothing else will still be expensive because the bigger cache may reduce yields. Even a microscopic defect in the cache die may cause it to be rejected during QA/QC. And an intact cache that large would still have higher latency than the smaller one it will be replacing. They need something like IBM's virtual cache implementation: https://www.anandtech.com/show/16924/did-ibm-just-preview-the-future-of-cachesWell, for one- I quadrupled the L2 cache from 12 MB to 48 MB.
Maybe X Plus with boost speeds disabledThe last Windows Dev Kit (Volterra) used the 8cx Gen 3.
This is very interesting. Has IBM patented this technology?Doing just that and nothing else will still be expensive because the bigger cache may reduce yields. Even a microscopic defect in the cache die may cause it to be rejected during QA/QC. And an intact cache that large would still have higher latency than the smaller one it will be replacing. They need something like IBM's virtual cache implementation: https://www.anandtech.com/show/16924/did-ibm-just-preview-the-future-of-caches
🚨🚨🚨 Heavy accusation from Charlie🚨🚨🚨
@FlameTail @adroc_thurston
Qualcomm Is Cheating On Their Snapdragon X Elite/Pro Benchmarks
Qualcomm is cheating on the Snapdragon X Plus/Elite benchmarks given to OEMs and the press.www.semiaccurate.com
Free article
Step forward a little in time. After OEMs got initial samples and made something close to the final designs, SemiAccurate got reports of poor performance. By poor we mean far sub-50% of the numbers Qualcomm was telling them in the technical docs and presentations.
Later, with more Snapdragon X Elite samples in the wild and many more revisions of WART, we got similar reports from OEMs and another Tier 1. Both reported numbers that were nowhere close to what Qualcomm promised. How not close? Above 50% this time but one used the term ‘Celeron’ to describe performance. The claims of better than Apple’s Rosetta 2 x86 emulation are clearly not real on what is probably the release hardware and software. Actually the silicon emulation may be better but everything else is unquestionably not.
Oh this is a gold mine...🚨🚨🚨 Heavy accusation from Charlie🚨🚨🚨
@FlameTail @adroc_thurston
Qualcomm Is Cheating On Their Snapdragon X Elite/Pro Benchmarks
Qualcomm is cheating on the Snapdragon X Plus/Elite benchmarks given to OEMs and the press.www.semiaccurate.com
Free article
I'm thinking someone gave some spoof info to him lol.Oh this is a gold mine...
That's good news: WoA machines will be a failure because of MS. Prices will go down. And Linux users who don't care how poor Windows is will be happy... provided one can install Linux that is🚨🚨🚨 Heavy accusation from Charlie🚨🚨🚨
@FlameTail @adroc_thurston
Qualcomm Is Cheating On Their Snapdragon X Elite/Pro Benchmarks
Qualcomm is cheating on the Snapdragon X Plus/Elite benchmarks given to OEMs and the press.www.semiaccurate.com
Free article
This is very interesting. Has IBM patented this technology?
This could be a viable hack to greatly boost ST performance of modern CPUs.
In X Elite for instance,
A single core can access it's private L1 and 12 MB shared L2.
With the above technique, a single core can access it's private L1, 12 MB shared L2, and a 24 MB virtual L3 (from 24 MB of L2 of the other two core clusters).
🚨🚨🚨 Heavy accusation from Charlie🚨🚨🚨
@FlameTail @adroc_thurston
Qualcomm Is Cheating On Their Snapdragon X Elite/Pro Benchmarks
Qualcomm is cheating on the Snapdragon X Plus/Elite benchmarks given to OEMs and the press.www.semiaccurate.com
Free article