• We’re currently investigating an issue related to the forum theme and styling that is impacting page layout and visual formatting. The problem has been identified, and we are actively working on a resolution. There is no impact to user data or functionality, this is strictly a front-end display issue. We’ll post an update once the fix has been deployed. Thanks for your patience while we get this sorted.

Discussion Intel current and future Lakes & Rapids thread

Page 87 - Seeking answers? Join the AnandTech community: where nearly half-a-million members share solutions and discuss the latest tech.
Pretty much goes to show that Ice Lake is a paper product. They can barely produce enough for two super-niche products, and meanwhile the bread and butter is still on the "well-aged" 14 nm variants (Whiskey, Coffee, Vodka...). The "improving" 10 nm yields are a slight of hand, and meanwhile we already have 7 nm products from the competitor...
 
The "improving" 10 nm yields are a slight of hand, and meanwhile we already have 7 nm products from the competitor...

I'm sure the yields are much better but still horrifically bad. Not sure if I totally believe him, but Charlie was saying that Cannonlake's yield was in the single digits % even accounting for the IGP being disabled which is like 2/3 of the chip. So this is an improvement in that they are actually able to ship enough for actual products even if it's only a very small fraction of sales.

My guess is that the i5 is either 40 or 48 EU btw.
 
GG, AMD. Went Zen and lost the GPU lead.

Isn't even the integrated Xe, nor is it Willowcove. Which are obviously going to be better than Sunny/G11.

So, while Intel goes revolutionary; Tigerlake is Xe. AMD is stuck in evolutionary design; Renoir is Vega.
 
Last edited:

Yeah I mentioned it a while back as almost assured, after it became clare that Icelake is using 3200 Mhz LPDDR4 (it's a shame that Picasso doesn't support that).
Worth to keep in mind that Vega is heavily BW bound as well as 2500U is actually faster than the 2700U you compared to.

Still AMD really needs Zen 2 and Na'Vi in mobile to be competitive with Icelake
 
Just when I thought I was out, they pull me back in!

Indeed.

Truth is that AMD's iGPUs haven't really improved much in . . . awhile. It's obviously no longer a priority for them. They don't care and/or don't have the resources to keep improving performance. Apple had them whipped months ago. Now it's Intel's turn.
 
need to know what machine the tested 2700u is in. If you follow the ryzen laptop market at all you'll know a lot of OEM offerings with raven ridge underperform for various silly reasons.

I'd be interested to see how the Gen 11 igpu compares to mobile vega when techepiphany gets his hands on one. Here's what the desktop vega APU can do when tweaked.
 
No, but it has much higher bandwidth. It could be as high as LPDDR4 4266.

Ah, OK. Well then, hopefully AMD put a lot of work in to there IMC and can reach speeds higher than 2400. At least 3200 should give it a huge boost, but obviously the more than squeeze out of it the better.
 
Ah, OK. Well then, hopefully AMD put a lot of work in to there IMC and can reach speeds higher than 2400. At least 3200 should give it a huge boost, but obviously the more than squeeze out of it the better.

I should add that I wouldn't expect the 64 EU model to be close to Vega 10, let alone faster in actual games, even with the bandwidth difference. Maybe in some titles close to Vega 8. I just don't think AMD bothers to optimize the drivers for Gfxbench at all compared to what is normally done for a title.
 
Just think, they could have stayed with Con cores and lost on both fronts!
I have thought about it. They most likely would have been ahead since 2014. If they kept the commitment to their earlier roadmap. Also, GlobalFoundries wouldn't have taken such huge losses in 2014/2015 to pay out Samsung.

They missed out on Posits/Valids in the late decade for GPUs. https://posithub.org/about
They missed out on the QuadFMA. AMD got AVX in only two quarters after Intel, and AMD could probably have implemented AVX512 in 2015~2016 on the exclusive 14nm GlobalFoundries process.
 
Back
Top