I can finally make this thread.
Phoenix is indeed RDNA3. My advice to everyone: treat the old APU rumours as being out of date.
Phoenix is indeed RDNA3. My advice to everyone: treat the old APU rumours as being out of date.
You won't have max power in laptops equipped with LPDDR5-6400 though as those machines are going to be a different class of products altogether. I suspect things will even out and essentially be the same in thin and light laptops (less power, more bandwidth vs more power, less bandwidth).Let's not forget that even through 780M was maxed out power wise, but It used only 4800MHz DDR5. 6400MHz offers 33% higher BW, so It should provide some FPS boost.
Maybe ComputerBase will do a comparison at different TDPs. Then we will see what clockspeed is achievable at lower TDP or how much performance is lost.You won't have max power in laptops equipped with LPDDR5-6400 though as those machines are going to be a different class of products altogether. I suspect things will even out and essentially be the same in thin and light laptops (less power, more bandwidth vs more power, less bandwidth).
At Notebookcheck.net they looked at Ryzen 9 7940HS.
View attachment 80320
Idle doesn't look better than the predecessor.
P.S. They will also look at the IGP 780M in a separate article.
You shouldn't tell me that. You should tell It to the author of that graph.The curve for the 13700H is not accurate and score is inflated, at 70W for the first run it score 15300 on the Samsung Galaxy.
You shouldn't tell me that. You should tell It to the author of that graph.
Theoretically, if your uncore power is high enough, but doesn't scale as rapidly, then of each additional watt, a greater percentage goes to the cores.It s for general information, one can notice that from 55W to 65W perf increase almost linearly with power, wich is just impossible
Of course, we were also interested in how the new Ryzen 9 7940HS performs with different power limits and have compared it to the Intel Core i7-13700H and the Asus TUF A15's AMD Ryzen 7 7735HS at 35, 45, 55, 65 and 80 watts. On both Asus systems, we were able to set power limits via the Armoury Crate software's manual mode - on the Intel system we used the Freeware tool ThrottleStop. We measured the following results:
The AIE is quite large according to one of their slides, and seemingly that's also cutCould they really chop off 40 mm2 with cutting just 2 cores and slashing the IGP?
It doesn't seem like it's using Zen 4c.
Saturated market = low prices... im not going to complain about that.
If this means we are going to be able to get cheap ARM, Intel and AMD handhelds that are petty much full pcs you can use in any way you want, i certanly like it.
(and this is when you realise how badly Intel screwed up by not having dual channel on ADL-N), this is the same mistake AMD did with Bobcat and and old small cores.
Theoretically, if your uncore power is high enough, but doesn't scale as rapidly, then of each additional watt, a greater percentage goes to the cores.
Regardless, they explicitly said they measured those numbers:
There are already some SBC or mini PC boards available like the udoo bolt, etc. They are mostly embedded SOCs so no changing the CPU...At least AMD is beginning to address the market.
I would love to see an AMD small form factor diy board + chip, however. I know we won’t see something the size of a raspberry pi, but still…
some more videos:There are already some SBC or mini PC boards available like the udoo bolt, etc. They are mostly embedded SOCs so no changing the CPU...
Links & videos:
For the U line sure, it's only an announcement but the HS laptops are already out, even only as ASUS models for now (TUF Gaming and ROG Zephirus 14 already available in Europe).Because it's only an announcement and another paper work from AMDs mobile lineup. AMD is so slow getting Phoenix out, looks like Phoenix laptops have to wait another few months.
For Windows.This software for handhelds
Not surprising then.Steam Deck: compatible but not fully supported
If I end up picking up an ROG Ally (and the chances are very much not 0) then I'm definitely going to be using this.Handheld Control Panel
This software for handhelds looks pretty nice.
Product support:
- Anbernic Win600
- Aya Neo: Original, Pro, Air, Air Plus, Next, Aya Neo 2, Aya Neo Geek.
- Ayn Loki
- GPD: Win 2, Win 3, Win 4, Win Max, Win Max refresh, Win Max 2, Pocket 3
- One X Player: Original AMD and intel, Mini AMD and Intel, Mini Pro, and the One X Player 2
- Steam Deck: compatible but not fully supported
Bobcat was a low end product for cheap and cheerful laptops, single channel was the right move. Dual channel would have increased the die size and pushed up platform costs, and most OEMs would only populate one channel anyway.(and this is when you realise how badly Intel screwed up by not having dual channel on ADL-N), this is the same mistake AMD did with Bobcat and and old small cores.
@Abwx was faster.
The best result was in CP2077, where It managed 26% higher FPS at Full HD high details over Yoga with 680M and similar power consumption. In some games, It's actually slower.
Pretty disappointing, faster RAM will help, but not so much.
For the U line sure, it's only an announcement but the HS laptops are already out, even only as ASUS models for now (TUF Gaming and ROG Zephirus 14 already available in Europe).
That RAM speed is only 4800 MHz, that's not enough even for Rembrandt!Proper comparison against 680M with same RAM:
~15% faster with +27% clock speed increase. Meh improvement, in my opinion.