Comments Locked

117 Comments

Back to Article

  • JfromImaginstuff - Tuesday, January 12, 2021 - link

    Seems pretty interesting
  • edgeofstar - Thursday, July 1, 2021 - link

    jjooooo
  • alufan - Tuesday, January 12, 2021 - link

    Erm RDNA2 is coming to mobile processors this year
  • Arsenica - Tuesday, January 12, 2021 - link

    They confirmed that they are launching mobile RDNA2 GPUs this year.

    In the APU front they haven't confirmed the release date for the Van Gogh (ZEN2 + RDNA2), Rembrandt or Dali? (ZEN3 + RDNA2) chips
  • Smell This - Tuesday, January 12, 2021 - link


    The discreet "Navi mobile" will be the 'next' APU graphics engine
  • Santoval - Tuesday, January 12, 2021 - link

    Yes, but only in lower end Van Gogh APUs which will only have ... Zen 2 APUs. In other words AMD wants to force their customers to pick either fast CPU cores or a fast iGPU but will not offer both in the same part until... 2022. Have they gone insane? Not even Intel have pulled that crap..
  • Santoval - Tuesday, January 12, 2021 - link

    edit : I meant Zen 2 *cores*.
  • dotjaz - Saturday, March 13, 2021 - link

    They are not forcing you anything. RDNA2 on APU can and will be slower than Vega in this case.
  • Spunjji - Wednesday, January 13, 2021 - link

    This kind of ventriloquising of large organisations is always a bit weird. Why on Earth would AMD *want* to "force" their customers to choose between a more advanced CPU and GPU if they could offer both?

    They likely have parallel design teams working on integration of the latest CPU and GPU cores. One project got one, one got the other, and they put them where they mattered most - area-optimised Vega for the Zen3 cores likely to be paired with dGPUs on a tight timeline, PPW-optimised RDNA2 for the low-TDP chips coming out later.

    The combination of both will arrive later, as has been the case with every previous integration of a new AMD CPU architecture with a new GPU architecture, but this time they're actually getting the cores into the new APU *way* earlier than they've managed before. That later combination might even benefit from DDR5, too, which would make a decently-sized RDNA2 iGPU actually worthwhile.

    As for "Not even Intel have pulled that crap" - see the Broadwell / Skylake transition. New high-performance cores or a large iGPU with a large cache - one or t'other.
  • Smell This - Wednesday, January 13, 2021 - link


    Actually, I like the choice __ as long as a graphics engine chiplet complete with Infinity Cache //FPGA// on die.
  • Smell This - Wednesday, January 13, 2021 - link


    . . . . in the **future**
  • HarryVoyager - Wednesday, January 13, 2021 - link

    I'm torn over whether a Zen3/RDNA2 APU would actually be interesting. They aren't going to have enough RDNA2 compute units to really be great gaming machines yet, and when they don't have the GPU grunt, they really don't get much out of Zen 3 either.

    My guess is they're focusing more on getting to enough bandwidth that they can put a PS5 on an APU, and looking more towards high effeciency CU's on the GPU.

    Now that said, these new high performance Zen 3 laptop chips, if they can get great efficiency on the Vega CU's could make very capable universal laptops when paired with the new nVidia mobile GPUs. Depending on how good their low demand efficiency is, I could see a gaming book that runs anything when plugged in, and can still have solid battery life for mobile, non-gaming use. Basically the sort of thing you'd load your Steam library on for gaming on the road.
  • Spunjji - Friday, January 15, 2021 - link

    At this point, I'm fairly convinced it wouldn't be all that great. The cost/performance side of things is always going to veer in favour of lower performance to keep costs down.

    That said, I still want one for my tiny ITX media box - something with roughly double the performance of the Vega 11 in the Ryzen 3400G ought to be achievable, and it would exceed PS4 levels of performance. For what I want to do with it, that's plenty. I'm happy to wait for now because there's literally no better way to get what I want (the box has a 90W PSU and no space for add-in cards).
  • GeoffreyA - Thursday, January 14, 2021 - link

    "parallel design teams"

    Agreed. Quite likely when the Cezanne team started work, RDNA2 wasn't ready, while Vega was pretty much "drop it in" and go. When the Z2 + RDNA2 team started, possibly Z3 wasn't complete (and RDNA2 as well). Even with the CPU architecture, there are at least two teams working in parallel; and in fact, we're lucky that Zen 3 actually inherited all of Zen 2's improvements Zen+. The fact that it did is testament to the excellent work going on behind their doors.
  • nandnandnand - Friday, January 15, 2021 - link

    It makes perfect sense when you realize you don't need a fast iGPU in a mobile chip that is going to be paired with a discrete GPU anyway.
  • Spunjji - Monday, January 18, 2021 - link

    👆
  • Kishoreshack - Tuesday, January 12, 2021 - link

    Was waiting to get a laptop
    I think Zen 3 would be worth the wait
  • marrakech - Thursday, January 14, 2021 - link

    it all depends on price, asus will have the new designed tuf a15 models with both an 4800h and an 5800h cpu
    id say for work the 4800H is enough no upgrade there and pricing will be favorable
    if the 5800h model will cost 300 dolar more since marketing hype over the 5 and 30 at the begining of the names id say its not really worth waiting for
    if you need the laptop to brag about 15 % faster single score sure go for the nev 5800h model
    if the prices vil differ 250 to 350 $ id get the 4800h models https://i.postimg.cc/TP4Kf9nn/Screenshot-204.png
    an good cpu is an good cpu
  • WaltC - Tuesday, January 12, 2021 - link

    Unless I was mistaken, Dr. Su also stated that AMD would be launching RDNA2 mobile processors later this quarter. What that means as to time to market availability is unknown, of course.

    Pretty straightforward, blasé presentation by Dr. Su, I thought. I hope that during this week AMD will address the significant product shortages it's had for a couple of months in terms of both Zen 3 CPUs and RDNA2 discrete GPUs. I miss the live presentations--these pre-recorded, heavily edited presentations are a bit too bland for my tastes...;)
  • The Hardcard - Tuesday, January 12, 2021 - link

    I got a 4800H which is fantastic however, I am tech lusting after that double double cache. I want to see what impact it has on video encoding and transcoding.

    However, I can resist until they start putting AMD in workstation laptops, ala Lenovo P Series s.
  • ajoy39 - Tuesday, January 12, 2021 - link

    Obviously its Zen2/4000 series but there are some P series models with AMD chips available right now
  • gijames1225 - Tuesday, January 12, 2021 - link

    So am I reading this correctly that the we now have hyperthreading in the U-series chips? That's pretty sweet if so. I'll be able to buy and ultrabook with 16 threads in a few months.
  • cfenton - Tuesday, January 12, 2021 - link

    The 4800u and 4600u had hyperthreading (SMT). They are just rarely used in OEM designs. For whatever reason, almost everyone used the 4700u and 4500u.
  • gijames1225 - Tuesday, January 12, 2021 - link

    Ah, thanks, yeah, I never even saw those ones when I was pricing a laptop a few months back. With the whole lineup getting SMT these will be really nice.
  • danbob999 - Tuesday, January 12, 2021 - link

    The problem with H series CPU is that you can't find them in non-gaming laptops.
    They have a perfectly fine GPU and could make great work machines.
  • gijames1225 - Tuesday, January 12, 2021 - link

    Yeah, I would love a slightly chunkier (like Thinkpad T-series thickness) laptop that was basically a workstation in a 14-in chasis. I just compile code all day, so I don't need more GPU than what is needed for a couple of monitors, but I'll devour all the RAM and CPU horsepower I can get.
  • jbwhite1999 - Tuesday, January 12, 2021 - link

    Check out the ThinkPad P15v or T15p - those have Intel (Sorry - no AMD) H-Class CPUs.
  • brucethemoose - Tuesday, January 12, 2021 - link

    Well the extra GPU in the G14 didnt come at much of a premium last year.

    Whether thats true for this gen remains to be seen.
  • Spunjji - Wednesday, January 13, 2021 - link

    Huawei released a Matebook 14 that uses the H series, but yeah, would be nice to see more designs like that.
  • ikjadoon - Tuesday, January 12, 2021 - link

    Gotta say, even with AMD's resurgence, M1 does take the shine off 40W, 50W x86 CPUs (or even TDPs, apparently now).

    I was really hoping AMD would have launched Jim Keller's K12 uarch, but we're stuck on x86 refreshes for how much longer?

    The mixing of Zen2 vs Zen3 is also, unfortunately, Intel-like and somehow it's even worse?
    At least Comet Lake vs Tiger Lake / Ice Lake had a different number of digits--how does Intel have less misleading model numbers versus AMD in 2020?

    Great to see the emphasis on battery life; they did quite well with Renoir, so let's please pummel Intel into the ground.

    But, I think I've now owned my last x86 laptop. :(
  • Bik - Tuesday, January 12, 2021 - link

    Very true. But apple is still apple, with their closed eco system and their steep price. We may have to wait until Arm design reach that level.
  • Exotic - Wednesday, January 13, 2021 - link

    The MacBook Air M1 price is not steep at all for an ultrabook, similar priced ultrabooks for $999 can't match the level of performance and battery life is what the MacBook Air offers now.

    For $999 the M1 Mac is great for students and edu pricing puts it at $899. Way better than Dell or HP or Lenovo at the same range
  • romrunning - Tuesday, January 12, 2021 - link

    re: Zen2 vs Zen3 - just remember for any 5000-series mobile, you should go for non-U processors first for peak performance. For battery life, if you get a U-series, only get the even-numbered ones.

    For the record, I hate the mixing in of Zen2 cores but keeping the 5000 numbering. I thought they were finally unifying everything under a common number, and then someone has to throw this in. Shame on whomever authorized that!
  • TheinsanegamerN - Tuesday, January 12, 2021 - link

    M1 offers 15W APU performance. How does that effectthe much more capable 40 watt chips in any way?
  • The Hardcard - Tuesday, January 12, 2021 - link

    outside of Taiwan being hit with a huge meteor or a Chinese invasion, it is clear that higher core Mxx SOCs are imminent. iIt is possible that there is demand for all, however any unsold machines will be the x86 ones, likely Intel far more affected than AMD.
  • Tams80 - Wednesday, January 13, 2021 - link

    We don't know how we'll Apple's slipcovers will scale though.

    I think it will be the best, but mainly due to them integrating much more on the SoC.

    And frankly, I'd rather sacrifice some performance on order to not have my RAM soldered on.
  • Kuhar - Wednesday, January 13, 2021 - link

    Very true. If you do ever-day tasks on your computer (word processing, web surfing, youtube, instagram, twitter etc.). When you need multicore performance, large amounts of RAM (and possibly combined with some more serious GPU), AMD owns Apple in every possible scenario.
  • Spunjji - Wednesday, January 13, 2021 - link

    "At least Comet Lake vs Tiger Lake / Ice Lake had a different number of digits--how does Intel have less misleading model numbers versus AMD in 2020?"

    Yeah, that's annoying. It's clearly so that AMD and OEMs can utilise unsold Renoir inventory. I guess it's the same old story - if you're the sort of consumer that doesn't already know the difference between Cezanne and Renoir, then you probably won't notice the performance difference between Zen 2 and Zen 3 as much as the price difference between the models.
  • Nicon0s - Thursday, January 14, 2021 - link

    I have to say these: no matter what AMD or Intel do Apple's M1 processor is still better" are quite amusing.
    I don't see how AMD would have faired better if they would have developed an ARM CPU, actually it doesn't make any sense to make such a claim. One of the main advantages of Zen processors is that they are an x86 arhitectue so they are fully compatible with the absolutely huge x86 software library.
    As a current Windows laptop user I see no attraction for an ARM MacBook other than the fact that is has good battery life. I honestly would never give up my software, games libraries to move to an ARM MacBook in order to get better battery life.
    People who believe that Windows laptops will massively lose sales in favor of these ARM MacBooks are simply delusional.
  • Ptosio - Saturday, January 16, 2021 - link

    Plus, as per article, Zen 3 is claiming 20h+ battery life.

    And all that on one generation of process node behind. Once AMD gets access to 5nm (and that's TSMS's technology, not Apple's), it's not impossible that they can achieve similar runtimes on x86.

    And the performance of mobile Ryzen is already more than adequate.

    I for one, don't see the appeal of ARM UNLESS Windows and Linux would have full x86 emulation that would run at least as good as the original.
  • 5j3rul3 - Tuesday, January 12, 2021 - link

    Zen 3 HS the Performence Monster
  • cfenton - Tuesday, January 12, 2021 - link

    The U series continuing to have Zen 2 is pretty dirty. I'm expecting OEMs to mostly offer the 5700u and 5500u, just like they mostly used the 4700u and 4500u. It's also weird that the 5600u will probably be better overall than the 5700u. It will be significantly faster single core, and probably about the same multi core.
  • shabby - Tuesday, January 12, 2021 - link

    Ugh that slimey as fuck, wth amd!
  • TheinsanegamerN - Tuesday, January 12, 2021 - link

    AMD has always been slimey. Same company was going to try and leave 300 and 400 series users high and dry with ryzen 5000, constantly delayed their mid range chipsets, killed the PCIe4 fromt he CPU with B450 despite there being no technical reason to not allow the option, And Ignored complaints about their poor quality video drivers until the tech press began slammign them (this has happened 3 times in the last decade). They jacked up the prices of their CPUs and GPUs the moment they had any sort of performance advantage.

    AMD is no better then intel or nvidia. The onyl reason they dont have as many shady dealings as the previosu two is simple lack of funds. IF they were making intel level money, it wouldnt surprise me to see the same tactics from them you see from intel and nvidia.
  • vladx - Tuesday, January 12, 2021 - link

    Don't forget about "Barcelona" in the 2000's.
  • Spunjji - Wednesday, January 13, 2021 - link

    Why pad out your rant with bullshit?
    "killed the PCIe4 fromt he CPU with B450 despite there being no technical reason to not allow the option" - the 400-series boards don't have the necessary traces and signal booster components for PCIe 4.

    It makes it even more obvious how weak the "they were going to do x... but didn't" argument is, because I cannot think of a single occasion when either Nvidia or Intel bowed to consumer pressure ob behalf of users of older hardware.

    Corporations are corporations, for sure. AMD would have to do a *lot* more shitty sutff to end up at the same "sliminess" scale as either Nvidia or Intel, though. Maybe they will now and this is just the start - who knows - but you threw in a bunch of equivocation and acted like you made a point here. You really didn't.
  • vladx - Tuesday, January 12, 2021 - link

    AMD always had misleading marketing, unlike Intel which only recently have done it.
  • Qasar - Wednesday, January 13, 2021 - link

    yea ok, sure, so says the pro intel fanboy vladx.
    amds marketing has been pretty accurate the last 2-3 years,while intels has lied every chance it could. it kept claiming that 10 nm was on track, is once such example
  • Spunjji - Wednesday, January 13, 2021 - link

    Intel were dishonest right back from the days of the first "Intel Inside" campaigns, where they sold mediocre hardware at premium prices off the back of marketing funded by fabulously high margins. They've been pushing FUD and misleading consumers for decades.

    I can never really tell whether vladx is an angry ignorant child, or just a liar.
  • Qasar - Wednesday, January 13, 2021 - link

    simple, vladx, is both
  • Spunjji - Wednesday, January 13, 2021 - link

    AMD's marketing since Zen has been 95% on-point.
    Intel have *literally never* had honest marketing, they just got away with it more when they weren't obviously screwing the pooch.

    You're a clown.
  • Tams80 - Wednesday, January 13, 2021 - link

    Is it?

    The U series APUs still have product names relative to their preformance, which is what most people care about, not the underlying tech.

    And the top end U series is Zen 3, so is it really much of an issue? If you care, you'll know and not buy the ones you don't want.
  • neblogai - Tuesday, January 12, 2021 - link

    Tech enthusiasts like to make a 'scandal' out of it , but really it will not make a difference for the majority of users. For example- games will run the same, as game performance will be bottlenecked by the iGPU anyway. For most (light) tasks- PC will feel the same, minute differences will not be noticeable in use. And multicore performance will also scale with SKU number (5500U<5600U<5700U<5800U)- so all good there too.
  • Spunjji - Wednesday, January 13, 2021 - link

    ^ This.

    I don't *like* it, but it's still better than having 2 entirely different product lines on different manufacturing nodes under the same name.
  • Tams80 - Wednesday, January 13, 2021 - link

    Precisely. As a bog-standard customer, if I bit a machine with a higher named SKU, then it will perform better than a lower one.

    The architecture doesn't really matter, as long as it delivers.
  • Hulk - Tuesday, January 12, 2021 - link

    I'll take a 5800U in my Surface Laptop 4 please.
  • AMDNeedsCompetition - Tuesday, January 12, 2021 - link

    Very disappointing. We get recycled Vega graphics, the architecture of which is now 3 years old, and Zen 3 is only coming to 2 of the SKUs. Truly lazy AMD. Obviously Intel is not keeping you on your toes.

    Especially disappointing considering AMD came up with the term APU, always pushed the envelope as far as integrated graphics were concerned, and now we get "it should be good enough for you". Anandtech I think you should push them harder on this.

    Is it because they want people to rather spend money on discrete RDNA2 GPUs? Even so, disappointing and lazy.
  • 5080 - Tuesday, January 12, 2021 - link

    Vega graphics is perfectly fine for this APU. The bottle neck is not the GPU, it's the DDR4 memory system. Navi wouldn't offer much more performance over the existing system. Navi will be available once they support DDR5 and a fast bus system in 2022.
  • AMDNeedsCompetition - Tuesday, January 12, 2021 - link

    Memory bandwidth is not the only thing that influences performance. As another poster has pointed out, this now means that Intel has the better mobile iGPU solutions. How embarrassing for AMD!
  • 5080 - Tuesday, January 12, 2021 - link

    Are you sure it's better? I haven't seen a direct comparison. As far as I can tell intel suffers from the same memory latency issues for iGPU performance as AMD does.
  • lmcd - Tuesday, January 12, 2021 - link

    Vega is mediocre independent of its memory problems. The iGP on Tiger Lake easily outpaces Vega.
  • Tams80 - Wednesday, January 13, 2021 - link

    Xe as an iGPU is better (as you'd body well expect being newer), but it's not that much better.

    It's in a weird spot where while it is more powerful, that extra power isn't that useful
  • vladx - Tuesday, January 12, 2021 - link

    Yes Intel Xe is better than Vega APU, from gaming to video decoding/encoding.
  • Spunjji - Wednesday, January 13, 2021 - link

    Xe on Tiger Lake is indeed faster than Vega 8 - though not as much as synthetic benchmarks might indicate, and with some exceptions due to driver bugs. It also has better video hardware.
  • vlad42 - Tuesday, January 12, 2021 - link

    And RDNA2 would have significatly helped to alleviate the memory bandwidth bottleneck thanks to the 128 MB Infinity Cache and using delte-E color space compression throughout the entire gpu (introduced with RDNA1). The delta-E compression update would allow more data to be stored on all sub-L2 caches. This would cause the gpu to have fewer cache misses and thus be more efficient with the system memory bandwidth it has. Vega, for comparison, only uses delta-E compression in the memory controllers and L2 cache.
  • mczak - Tuesday, January 12, 2021 - link

    There's no way a rdna2 based APU would have 128MB Infinity Cache - the cache alone takes up over 100mm² die space, which would be completely unacceptable for such an apu. You'd have to scale that back a lot (probably to 16MB or so) to be viable, and I'm not sure how much it would help at that point. (Maybe future AMD APUs will actually have shared cpu+gpu cache, intel does this since ages, ARM SoCs have such caches (called system level cache there) too, right now with how AMDs cpu L3 cache is organized that's not really possible, since the L3 is a victim cache to the cores within the CCX cluster, so it would need to be an additional cache level.)
    You are right though that rdna2 (and rdna1) should be more bandwidth efficient than vega, even without additional cache, due to the changes in how framebuffer compression is handled.
  • TheinsanegamerN - Tuesday, January 12, 2021 - link

    Why not? Intel did a APU with broadwell, that had 128MB of cache, on 14nm, and it worked great. Even their cut down 64MB iris pro models had notable performance improvements.
  • mczak - Tuesday, January 12, 2021 - link

    The broadwell APUs were using a off-die L4 cache - this was using a separate 128MB edram die, built on a 22nm process, which measured "only" 77mm². Note you cannot integrate dram into the "standard" 7nm TSMC process, and rdna2 infinity cache is just ordinary sram integrated into the same silicon as the gpu itself, hence why it needs more die space even though it's using a much smaller manufacturing process.
    And this 7nm TSMC process is expensive, and availability is actually scarce, so 128MB is definitely not an option (unless AMD would try off-die cache as intel did, not sure how viable it would be nowadays).
  • Spunjji - Wednesday, January 13, 2021 - link

    There's a reason Intel only did that once, and charged an absolute fortune for it when they did.
  • vlad42 - Tuesday, January 12, 2021 - link

    I agree that a 128 MB cache would not make sense at least until stacking is viable (5nm perhaps based on articles from this sight covering TSMC?). However, I think a 64 MB or 32MB cache might be reasonable. Picasso is 209.78 mm2 compared to Renoir which is 149.22 mm2. So, with 64 MB of cache, that could still make for a smaller chip than the 12nm Picasso depending on the die size increase with Cezanne. If that ends up being too big due to manufacturing and pricing constraints (AMD could always have a 2 APUs, one high cost and one low cost much like Intel does), then 32 MB should be doable as I imagine it would probably be pretty close to what the 3 CU difference between Picasso and Renoir would amount to - maybe the cache would be a little larger. I will admit, I have not been able to find out how big the 7nm Vega CUs are. That said, assuming I am interpenetrating the Big Navi die shot correctly, it looks like the size of 32MB of Infinity Cache is between the size of 3 and 4 of the RDNA2 CUs.

    As for benefits, the 32MB edram cache did wonders for the original Xbox One and that cache had far less bandwidth and no color space compression (so it would be effectively smaller). I have no idea how beneficial a 16MB cache would be but I doubt the scaling is linear - meaning I think it will provide worse than linear scaling at those capacities. However, relying solely on DDR5 is a bad idea. It looks like AMD, Intel, etc., will jump straight to the fastest JEDEC spec for DDR5 (6400) or very close to the fastest. While this will be a big boost to performance when it is first adopted, there will not be significant improvements again until DDR6, or whatever replaces DDR5. This will still limit the maximum bandwidth available to the iGPUs to roughly 64-72 GB/s when we take into account the bandwidth needed for the CPUs. This would puts the iGPU between the MX350 and MX450, old low end GPUs by that point, in terms of raw bandwidth and behind both once Nvidia's superior color space compression is considered - I have not heard of AMD increasing the amount of compression since Polaris, just expanding where the compression is used. This will be very limiting long term because it will be years before a replacement for DDR5 is available.

    While it would be interesting if AMD were to use the CPU's L3 cache instead of a dedicated Infinity Cache, I doubt they would want the CPUs in APUs to have access to more L3 cache than the dedicated CPUs seen in the desktop Ryzen, Threadripper and Epyc chips. I also suspect that the L3 cache is optimized far more for latency over bandwidth, which would be to the detriment of the iGPU both from a bandwidth and capacity standpoint. I could see AMD introducing an L4 cache though. This could then be put on the IO die for the CPU only parts and avoid the problem of the high performing CPU only parts having access to less cache.
  • vladx - Tuesday, January 12, 2021 - link

    Well said, AMD would just release rehash after rehash without competition, at least Intel had an excuse with their 10nm failures.
  • The Hardcard - Tuesday, January 12, 2021 - link

    What? How about AMD did all of this with almost no money. Their entire lineup had been bargain basement and undesired for years. they could have just as easily have gone out of business as had this success, it was 50-50.

    They did the right thing by focusing on basic compute first. They now have a number of products that people will buy and can use. Obviously now, they have a lot of work to do in many other areas. While they have achieved top line the line traditional CPU and GPU compute, there are a lot of reasons for people who are looking toward the future of computing to buy Intel and Nvidia.

    The Vega issue was simply because Dr. Su made the choice to advance the CPU side first. For the prospects of AMD that was the right choice. So the market has to wait for them to catch up otherwise. There still millions of people who can use a good old fashion simple powerful CPU.
  • Qasar - Wednesday, January 13, 2021 - link

    The Hardcard, ignore vladx, he loves intel to no end.
  • Spunjji - Wednesday, January 13, 2021 - link

    Oh cool, yet another single-issue rant account. Yay.

    "Zen 3 is only coming to 2 of the SKUs"
    Someone only read the bit of the article they wanted to whinge about.
  • dodoei - Tuesday, January 12, 2021 - link

    In the meantime, 4800Us are still not available in the US
  • lightningz71 - Tuesday, January 12, 2021 - link

    Considering that the biggest hurdle to mobile APU gaming is on the gpu side, it's going to be interesting to see how the benchmarks shake out between the 5600u and the 5700u. Both are specced for the same DDR speeds. On the one hand, the 5700u has one additional CU running a few hundred MHz faster. On the other, you have the much larger L3 cache and unified pool setup. On the CPU side, the 5600u looses a bit on max boost MHz, but does much better per clock, whereas the 5700 boosts higher and has more physical cores.

    This should be interesting.
  • 0iron - Tuesday, January 12, 2021 - link

    It's a weird time now. We have an Intel iGPU better than AMD! Who would have thought that even 2 years ago?
  • Bik - Tuesday, January 12, 2021 - link

    Dr Su said time to market didn't align to use rdna on ryzen 4000, makes it look like redemption will come with ryzen 5000. Nope they just didn't care. The mobile team just swap zen 2 on renoir with zen 3 compute die and call it a day. Super disapointing.
  • AMDNeedsCompetition - Tuesday, January 12, 2021 - link

    Nope, they didn't care and were too lazy. Might put off my upgrade from a 3500U based laptop until AMD releases a proper 6000 series. Honestly, the knowledge that I'd be buying the same GPU architecture as my 3500U has is very offputting.
  • Smell This - Tuesday, January 12, 2021 - link


    You not understand how AMD rolls. CPU engine first ---> THEN graphics engine. It has always been that way with APUs. That is why the Ryzen 5000 Mobile series runs with a Vega graphics engine.

    The 3xxxU and 4xxxU mobiles were the 'blip' --- the CPU engines went Zen+ to Zen2.

    With the next round Ryzen 5000s / Zen3, I suspect a 'Little Navi' mobile graphics engine with a 'puny' 12-16 CUs and Infinity Cache. The 'mobile' Xe MAX graphics will be quaking in its boots. HA!
  • Spunjji - Wednesday, January 13, 2021 - link

    Same reason here - time to market. This is the fastest an APU has followed a new CPU architecture release in AMD's history.

    But sure, they "don't care". That definitely sounds like the motivation for a multinational corporation.
  • lemurbutton - Tuesday, January 12, 2021 - link

    M1 is significantly better.
  • 5080 - Tuesday, January 12, 2021 - link

    If true, it doesn't help 92% of notebook users without an OS that they can use.
  • 1_rick - Tuesday, January 12, 2021 - link

    Go buy a Mac, then.
  • SaolDan - Tuesday, January 12, 2021 - link

    Yuck
  • 1_rick - Tuesday, January 12, 2021 - link

    Only way to get an M1, though, so the M1 is useless for non-Mac people.

    I have no reason to hate a good ARM Windows machine, when they come around, although if the software I use at my day job doesn't get ported or if the emulation isn't good enough, I wouldn't want to use one at work.
  • mdriftmeyer - Tuesday, January 12, 2021 - link

    It’s not wrt single core and the M1 gets spanked I. Multi core loads
  • GeoffreyA - Thursday, January 14, 2021 - link

    "M1 is significantly better"

    Renoir was right on this thing's tail, so Cezanne ought to cover the ground, while being produced at 7 nm too.
  • Hul8 - Tuesday, January 12, 2021 - link

    A note on "enabling each processor to have access to all the cache on the CPU *at the same time*" (emphasis mine):

    If these mobile Zen 3 cores still have a 2048-entry translation lookaside buffer(*, with 4KiB pages each core can only address maximum of 8MB of L3 at the same time.

    -

    *) as the desktop CPUs have; see pg. 4 of the article "AMD Zen 3 Ryzen Deep Dive Review: 5950X, 5900X, 5800X and 5600X Tested"
  • Santoval - Tuesday, January 12, 2021 - link

    "We’re at a stage now where AMD might consider updating the CPU/GPU on its APUs in alternate years, if that keeps the rate of product releases in line with its other designs."

    AMD have both retained an obsolete Vega iGPU for a *fourth* consecutive APU release (2000, 3000, 4000 and now 5000 series) and to add insult to injury, they bizarrely chose to rebadge 3 out of the 5 -U parts. It is not "alternate years" if for 4 years they keep the same iGPU; if it was alternate years, since Vega was introduced with the 2000 series, the 4000 series last year would have had a new iGPU and the next new iGPU would be introduced with the 6000 series next year.

    On top of that they are retaining PCIe 3.0 and (LP)DDR4 support when Intel have already added PCIe 4.0 support to Tiger Lake and it also has provisional LPDDR5 support (though not yet materialized). Last but not least their Van Gogh APUs later *will* have RDNA2 iGPUs and LPDDR5 support (maybe even LPDDR5?) but will only be paired with ... Zen 2 cores.

    Since Tiger Lake's largest Xe iGPU variants can decimate AMD's obsolete Vega iGPUs have AMD decided to offer back to Intel whatever little share of the mobile market they have gained as a gift?
  • Spunjji - Wednesday, January 13, 2021 - link

    "obsolete Vega iGPU"
    Intel only just got around to producing something that beats it, so it's hardly obsolete. AMD have improved Vega's performance per area *and* watt with every update.

    "Since Tiger Lake's largest Xe iGPU variants can decimate AMD's obsolete Vega iGPUs..."
    Check out the performance in actual games where the frame-rates are playable:
    https://www.notebookcheck.net/Vega-8-R4000-vs-Iris...
    They trade blows, with Intel averaging around 10% advantage. Literal decimation, I guess? But woo, check out those wins in synthetic benchmarks, it's 1997 again 😅

    If iGPU performance were so important then AMD would have led Intel for quite a while now. It's part of the story, but not the whole story, and the truth is that these solutions are *competitive*.
  • wrkingclass_hero - Wednesday, January 13, 2021 - link

    Well dang. I was hoping for RDNA 2 in the U series chips that would eventually make its was into a GPD Win device, but I might as well get the Intel based Win 3 since AMD obviously does not care about these chips.
  • s.yu - Wednesday, January 13, 2021 - link

    >appease OEM partners that have a singular design and want to get the benefit of the latest generation nomenclature but not have the expense of developing a new unit

    That makes me want to puke.
  • Rassassin - Wednesday, January 13, 2021 - link

    Why oh why is AMD ignoring PCIE 4.0 support for laptops. When RTX IO/DirectStorage support comes, this will seem a major oversight.
  • Spunjji - Wednesday, January 13, 2021 - link

    As always it's wonderful to see that while the Intel articles have attracted relatively few comments, the AMD articles have - as per usual - been astroturfed with FUD. Hooray!

    It's a shame, as there are things I'm genuinely disappointed by - I'd have preferred not to see Vega again, and the recycling of Renoir is a bit crappy - but I've just wasted a bunch of my time pointing out to single-issue shills that their SKY IS FALLING ranting is completely overblown, and now I now lack the energy to bother with nuance. So instead, let's go with:

    INTEL WILL BE CRUSHED, ZEN 3 FTW, AMD FOREVER
    I WILL BUY 20 OF THEM
    RAWR

    Huzzah, balance is restored to the comment section. 😑
  • GeoffreyA - Thursday, January 14, 2021 - link

    Thanks for the effort, as always. Well, I'm a bit disappointed myself about Vega but was expecting it, so that's that. RDNA2 in Cezanne would've been grand, certainly. Why, AMD, why :) Also, their mixing Zen 2 and 3 under the 5000U range is a bit concerning, and I just hope they don't continue this practice. Other than that, Cezanne is an excellent product and will cause bloodshed when it hits the benchmarks.
  • SolarBear28 - Wednesday, January 13, 2021 - link

    It is disappointing yet sadly understandable that Zen 2 cores are used for some of the U-series parts. The smaller cache probably make these slightly cheaper to manufacture. Its just not worth AMD's resources when OEM's mostly want either gaming machines or budget thin and lights. AMD delivered the parts (at the prices) OEM's will buy. At least there are some Zen 3 U-series parts for premium laptops.

    Unfortunately retaining the Vega iGPU was also the smart choice. Integrating RDNA 2 would probably add at least 6 months of delay if not more, and who knows how much it would cost. Its much more important from a financial perspective for AMD to maintain their momentum (and gaming CPU lead) and get the products out. Now they have class leading CPU's to pair with soon to be released mobile dGPU's from Nvidia and AMD.

    Integrating a desktop graphics architecture into a mobile monolithic die (and doing it well) is a tonne of work. Look at how long it took to create a Vega APU with good battery life. Look at how infrequently Intel introduces a new iGPU architecture. I think it was extremely smart of AMD to push out Zen 3 mobile parts quickly. Their market share is continuing to grow and there are more AMD options in laptops than ever before. (I'm typing this on a Thinkpad T14s with Ryzen 7 Pro 4750u and the combination of performance, battery life and portability is fantastic). Yes, it would have been nice for the fans if AMD maintained their iGPU lead. But that will be difficult to do if AMD continues to include more CPU cores than Intel, there's just not enough die space. Despite this disadvantage AMD is still able to offer better iGPU's in budget laptops.
  • Tams80 - Wednesday, January 13, 2021 - link

    This.

    I'm surprised, but happy that AMD have released new mobile CPUs so soon after the desktop ones and even relatively soon after the last gen.

    And while good iGPUs are great to have, the Vega is still pretty good, good enough for an iGPU. Especially on the higher TDP APUs, the iGPU isn't that important. A great one would be nice, but it isn't needed. And the thin-and-light iGPU only gaming market is quite small. None of which would have been worth sacrificing the launch of the 5000 series to six+ months later.

    OEMs were testing the water with the 4000 series. I doubt they would have been to happy to wait until summer to announce their new products, and would have just gone with Intel, leaving AMD out yet again. Instead, AMD are set to have an almost equal position with Intel, which is frankly, astounding considering where they were just a few years ago.
  • Spunjji - Friday, January 15, 2021 - link

    The thin-and-light gaming market is currently populated by designs with low-end dGPUs anyway, so designing an APU to compete in a world where LPDDR5 is still absent would result in a halfway-house compromise that's significantly slower than the competitors without offering any particularly compelling advantages. Combining a U-series with a basic iGPU and a low-end dGPU makes more sense right now.
  • GeoffreyA - Thursday, January 14, 2021 - link

    Agreed.
  • Spunjji - Friday, January 15, 2021 - link

    Nice to see a balanced opinion! I suspect we'll see AMD retake the iGPU crown back when they can combine the area benefits of 5nm with DDR5 - they won't need to use the area for more CPU cores this time around, so they could squeeze in more RDNA2 ALUs and maybe even a nice big cache for it.
  • marrakech - Thursday, January 14, 2021 - link

    so geekbench is an good benchmark?
    Ryzen 9 5900HX that was seen on Geekbench 5.3 racing at 4.59 GHz and producing a mighty single-core result of 1,423 points but a muted multi-core result of 6,912 points. Well it seems things have been properly tweaked, as Tum Apisak has found an Asus ROG Zephyrus Duo 15 SE (GX551QS) machine soaring at 4.74 GHz and coughing up a still great single-core score of 1,547 points and a huge multi-core score of 9,069 points
    6900 to 9000 is some tweaking , either one was run on eco mode or what is going on?
    id say an benchmark that has 2000 points diffrence on same cpu is abit scewed
  • SolarBear28 - Thursday, January 14, 2021 - link

    The TDP of the Ryzen 9 5900HX can be configured from 35-54 W. Plus all of the usual power mode, OEM firmware and thermal solution variables that could affect the results.
  • Spunjji - Friday, January 15, 2021 - link

    This - the dramatic difference between the two scores suggests the first device was hitting thermal limits will all 8 cores loaded. This is the problem with notebook SoC performance links - we have no idea about cooling!
  • marrakech - Saturday, January 23, 2021 - link

    https://browser.geekbench.com/v5/cpu/6041506 17 ROG Strix
    same laptop is leader in passmark cpu test 25000 (this goes over 60000 for treadripper parts)
    ... https://i.postimg.cc/PxvrnVrL/Screenshot-576.png
    on the left is the rog strix compared to an 1 year old 4800h part
    so in one banch its on top in other its very low?
  • ECC_or_GTFO - Friday, January 15, 2021 - link

    Can one of you with connections please contact the laptop makers and get them to support ECC on the motherboard? It doesn't have to come with it - but just support it; the CPU does! It would be cheap & huge with the enthusiast - preferably Thinkpad.

    Also get them to hire a guy like this for a friggin month to tune it so we can have low power like Apple's M1: https://www.reddit.com/r/hardware/comments/k3iobs/...

    Thanks! Such a paint to get someone's ear at these hardware companies these days... Despite working in "silicon valley" for a decade...
  • Meteor2 - Saturday, January 16, 2021 - link

    Cannot believe they're doing an Intel on the U-series. Absolute stupidity.
  • Gondalf - Saturday, January 16, 2021 - link

    No, unfortunately in U serie they are in a bad situation. 8 cores do not fit in 15W and Zen 3 is notoriously a lil power hungry for 7nm. So they have few low power good dies with Zen 3 and many low power good dies on Zen 2. It is a matter of binning. They are constrained to stay on Zen 2 for power consumption reasons, funny enough in the cpus that will sell more, aka 5700 U, 5500U, 5300 U
    One this is certain with the arrive of Tiger Lake 8 cores, the landscape will change radically.
    Anyway, Amd have not enough silicon, so to Intel care nothing of Amd claims.

    There is another issue, the battery life, Amd laptops are a delusion in this champ, the reason is AMD U APUs are 25W and definitively not 15 W. Amd is not fair here.
    Actual real TDP of Amd APUs are 25W, 35W,45W, 45W+++.
  • SolarBear28 - Sunday, January 17, 2021 - link

    That is a giant load of imaginary crap. Zen 3 power consumption is almost identical to Zen 2 at the same frequency. With the 19% IPC increase Zen 3 will be even more efficient than Zen 2 in real world usage in laptops. You say 8 cores do not fit in 15W? AMD has already done it. I'm typing on a T14s with Ryzen 7 Pro 4750U. Great battery life. Better sustained performance and efficiency under load than any product Intel has ever made for thin laptops. 25W PL2, 18W PL1 yet destroys the Intel i5-10210u (in the same laptop with the same power limits) and keeps up with a 45W i7-10875H. The only truth in your comment is that AMD does not yet have the production capacity to realize the market share their CPU's would otherwise have.
    https://www.notebookcheck.net/Lenovo-ThinkPad-T14s...
  • Smell This - Sunday, January 17, 2021 - link


    No worries ___ Gondaft is all about the propaganda.

    Ian is busy enough, but a blow-by-blow deep dive (complete with die shots!) would answer multiple questions between AMD and Chipzillah, including actual power consumption and "core" (both CPU and graphic engine) efficiencies.

    I'm thinking 'Daftie would likely not enjoy Vega 8 CUs versus Xe 96 EUs in any metric, much less with an incoming Little Navi iGPU.
  • Qasar - Sunday, January 17, 2021 - link

    " No worries ___ Gondaft is all about the propaganda."
    as well as bashing AMD any chance he can, and trying to make his beloved intel, look better then they are.
  • Spunjji - Monday, January 18, 2021 - link

    It's a pretty sad comparison. Intel needed around 33% more area to get themselves an average 15% performance advantage in actual games, and performance-per-watt is a similar story.

    It's a huge improvement over their previous efforts, but I don't find it to be very encouraging for their incoming discrete products.
  • Spunjji - Monday, January 18, 2021 - link

    "One this is certain with the arrive of Tiger Lake 8 cores, the landscape will change radically."
    How so? They'll all be a nominal 45W TDP (likely ~80W boost), so they'll be competing with the Zen 3 HS chips with higher IPC and better performance-per-watt. They're also late, so they're not likely to be making much of a change to the gaming notebook market.

    D'oh! 😂

    Really enjoyed your waffle about how AMD are supposedly over their TDP while not mentioning Tiger Lake at all. Very cute, classic Godnalf.
  • healthyCompetition - Monday, January 18, 2021 - link

    Whilst it's disappointing to see Zen2 chips in the 5000 U lineup at least they've changed them a little. The 4300u (4/4), 4500u (6/6) and 4700u (8/8) previously had no additional threads over the # of cpu cores.
    These 5000U chips all have double the threads (All are 'hyper threading' enabled).
    Great for those that want a little more grunt from the same chip.
  • six_tymes - Wednesday, January 20, 2021 - link

    after decades, amd PR team still cant come up with a better photo session. they still direct their ceo to stand still holding a processor up near eye level while wearing those silly outdated wrap around head mic. That is an old INTEL image done decades ago. you would think with all AMD's new found saudi money they would hire a more creative photographer and better PR team. Do something new amd.
  • Smell This - Wednesday, January 20, 2021 - link


    Yeah ... Poor ol' AMD. Chipzillah is 'eating their lunch' and 'spanking their fannies'
    ( eye roll )

    A few years ago I had a girl friend. She had been married 3 times and was still a virgin. As we got close I asked her about all her husbands. She said her first husband was a psychiatrist, and all he wanted to do talk about it. Husband #2 was an OB/GYN, and all he wanted to do was look at it. I asked about her third husband and she said he was an Intel fan, and all he would say is . . .

    "Wait until next year."
  • Mdarrish - Wednesday, February 10, 2021 - link

    Asshole.

Log in

Don't have an account? Sign up now