You really are throwing everything including the kitchen sink at generating FUD here. The only one of your numbers that represent reality is the 40% better power efficiency - the rest are clearly not comparing apples to apples, as the average ST/MT performance difference between Ice Lake and Zen+ difference is closer to 10-15%. That's what people mean when they say "competitive" (especially considering cost).
Tiger Lake is also most likely not coming "this summer". It's H2 2020 in Intel speak, and given Ice Lake was ~3 months later than they said it would be, I'm not prepared to assume that's anywhere near "Summer".
Yes you are. That article is exactly where I went to check the data *before I wrote my first reply*. The only specific ST test quoted has a 20% advantage to Intel, and it's in Cinebench, which is apparently Not A Real Benchmark now that AMD are predicting wins in it. So that's your one factual claim cut in half, and still nothing to back up your speculation.
The biggest wins for Intel in the article you're mentioning are in Handbrake (MT, not ST) and 7zip (unsure). 7zip is strongly affected by memory bandwidth, which is an advantage that Renoir negates by supporting LPDDR4x. The other big wins for Intel were in the web browsing tests - Renoir's changes to boost behaviour should help there, too, though we don't yet know by how much. What's absolutely certain is that none of those performance gaps will change in Intel's favour.
It's impossible to gaslight people when they have access to the same information you do - you just come out of it looking like a troll.
Dude, You really don't know how to read, that is the best answer to you. It writes like this after spec benchmark: with the Intel variant of the Surface Laptop 3 being ahead by 37% in the integer suite, and 46% in the floating-point suite. What is so hard to believe? Zen 2 is something like 15-20% better than Zen+. Ice lake is another easy 10% over Zen 2 in ipc. Add that and you will get the figure the dude you are arguing about said.
Depends if they done the bench suite at 25W or 15W. With last Ryzem mobile they were smart enough to test at 25W to show strange things. Anyway my bet with all cpus active they will destroy the battery with an huge power consumption utilizing the Tskin methodology to cool the SKU. It is now all gold under the sun. The idle power will be huge and the SOC availability small. No volume no party.
You will have bad surprises when someone will test these 8 core SKUs on a real laptop constrained to 15W. Knowing to be unable to beat Intel at the same number of cores, AMD chose another sad street. They claimed great things with present Ryzen Mobile, but real world power measures say the sad truth. Basically they not even try to ship a good low idle power four core eight threads, the most wanted in tiny devices.
We’ve updated our terms. By continuing to use the site and/or by logging into your account, you agree to the Site’s updated Terms of Use and Privacy Policy.
Facebook
Twitter
RSS
406 Comments
View All Comments
Nicon0s - Tuesday, January 7, 2020 - link
Multi-core was like 11%These new AMD APU's have 2X the cores so AMD is now winning in multi-core by a lot.
Spunjji - Tuesday, January 7, 2020 - link
You really are throwing everything including the kitchen sink at generating FUD here. The only one of your numbers that represent reality is the 40% better power efficiency - the rest are clearly not comparing apples to apples, as the average ST/MT performance difference between Ice Lake and Zen+ difference is closer to 10-15%. That's what people mean when they say "competitive" (especially considering cost).Tiger Lake is also most likely not coming "this summer". It's H2 2020 in Intel speak, and given Ice Lake was ~3 months later than they said it would be, I'm not prepared to assume that's anywhere near "Summer".
generalako - Wednesday, January 8, 2020 - link
No I'm not. Go check the Anandtech Ice Lake vs Picasso Surface Laptop 3 review. The single-core performance was 40% above it.Spunjji - Wednesday, January 8, 2020 - link
Yes you are. That article is exactly where I went to check the data *before I wrote my first reply*. The only specific ST test quoted has a 20% advantage to Intel, and it's in Cinebench, which is apparently Not A Real Benchmark now that AMD are predicting wins in it. So that's your one factual claim cut in half, and still nothing to back up your speculation.The biggest wins for Intel in the article you're mentioning are in Handbrake (MT, not ST) and 7zip (unsure). 7zip is strongly affected by memory bandwidth, which is an advantage that Renoir negates by supporting LPDDR4x. The other big wins for Intel were in the web browsing tests - Renoir's changes to boost behaviour should help there, too, though we don't yet know by how much. What's absolutely certain is that none of those performance gaps will change in Intel's favour.
It's impossible to gaslight people when they have access to the same information you do - you just come out of it looking like a troll.
yeeeeman - Sunday, January 12, 2020 - link
Dude, You really don't know how to read, that is the best answer to you. It writes like this after spec benchmark: with the Intel variant of the Surface Laptop 3 being ahead by 37% in the integer suite, and 46% in the floating-point suite.What is so hard to believe? Zen 2 is something like 15-20% better than Zen+. Ice lake is another easy 10% over Zen 2 in ipc. Add that and you will get the figure the dude you are arguing about said.
Korguz - Sunday, January 12, 2020 - link
but are the freqencies the same ?? seems.. intel cant reach higher clocks on there 10nm process.. that still true ?Gondalf - Tuesday, January 7, 2020 - link
Depends if they done the bench suite at 25W or 15W. With last Ryzem mobile they were smart enough to test at 25W to show strange things.Anyway my bet with all cpus active they will destroy the battery with an huge power consumption utilizing the Tskin methodology to cool the SKU.
It is now all gold under the sun.
The idle power will be huge and the SOC availability small. No volume no party.
Spunjji - Tuesday, January 7, 2020 - link
Gondalf reliably providing a useful guide on what not to believe, as ever.Gondalf - Tuesday, January 7, 2020 - link
You will have bad surprises when someone will test these 8 core SKUs on a real laptop constrained to 15W. Knowing to be unable to beat Intel at the same number of cores, AMD chose another sad street. They claimed great things with present Ryzen Mobile, but real world power measures say the sad truth.Basically they not even try to ship a good low idle power four core eight threads, the most wanted in tiny devices.
Jugotta Bichokink - Tuesday, January 7, 2020 - link
Gondalf your speculations are as useless as Intel's own.