Comments Locked

105 Comments

Back to Article

  • jeremyshaw - Friday, January 10, 2020 - link

    If the 15W chip can TDPup to 25W, can the 28W chip do something similar?
  • s.yu - Friday, January 10, 2020 - link

    It doesn't work like that, the TDP is the power envelope at base clock, it's the same chip (though perhaps with better binning) with a higher base clock, as is pretty clear from the chart.
  • Retycint - Friday, January 10, 2020 - link

    Technically the TDP isn't the power envelope at base clock, but rather the base clock is the minimum guaranteed clock speed at the TDP level. So certain workloads may see a higher clock speed at the same TDP level
  • Tigersam - Saturday, January 11, 2020 - link

    Yes, you are right. Base clock is just the min clocks it will run. Higher speeds are dependent on cooling and power limits.
  • Valantar - Saturday, January 11, 2020 - link

    *except for AVX workloads and when the iGPU is under load simultaneously.
  • Tigersam - Saturday, January 11, 2020 - link

    Yes, Intel has a short term boost that can go 40-50W..... But long term power is limited to 28W.
  • Gondalf - Saturday, January 11, 2020 - link

    So it is pretty safe to assume that OEMs are not interested much in a 28W SKU. They mainly want a cpu with a pretty stable (after few seconds) 15W power consumption (chispet inlcuded) to enlarge the life of Laptop batteries.
    It is not a casuality that early Ryzen Mobile 4000 laptops are relegated to a small 25W nice done of so so laptops made for some casual game without a powerful discrete GPU.
    My suspect AMD choses the 8 core way knowing they are unable to compete in the 4 cores 8 threads arena that is by far the largest one thanks to a very low idle power. When the new Ryzen will are out in some volume Tiger Lake will be in volume too, they will are absolutely unable to make a dent at that time.
  • SolarBear28 - Saturday, January 11, 2020 - link

    Tiger Lake is many many months away.
  • arashi - Saturday, January 11, 2020 - link

    Assuming HSteward got banned and came back as Gondaft. Next is F_____?
  • flgt - Saturday, January 11, 2020 - link

    OK, this was pretty funny.
  • JKflipflop98 - Sunday, January 12, 2020 - link

    Gondaft the dumb. Yes, that's what they called me.

    I'm Gondaft the Stupid. And I've come back to fanboy for AMD.
  • close - Monday, January 13, 2020 - link

    They're actually fanboying for Intel ;). Know your troll.
  • Techtree101 - Saturday, January 11, 2020 - link

    End of this year as announced by Intel (but we know how those announcements go).
  • Irata - Saturday, January 11, 2020 - link

    You do know that Ryzen 4000 mobile exists in both 15W (U series, can be configured up to 25W) and 45W (H series)?

    So not sure what you are on about with your „25W niche“ drivel.
  • Korguz - Saturday, January 11, 2020 - link

    its gondalf.. he will always make amd look bad.. while praising his beloved intel
  • Alexvrb - Sunday, January 12, 2020 - link

    *make himself look bad

    You had a typo.
  • Korguz - Sunday, January 12, 2020 - link

    nope.. no typo.. he wants to make amd look bad..
  • hecksagon - Saturday, January 11, 2020 - link

    Much easier explanation. AMD knows Intel can't compete in the 8 core market.
  • oleyska - Sunday, January 12, 2020 - link

    Much better explaination.
    Intel will come with tiger.
    8 core will dominate until that time and compete with tiger, or still win over tiger.
    6/12 wouldn't do it, it's a year till next apu release.
  • Korguz - Sunday, January 12, 2020 - link

    when will intel have more then 4 cores from its 10nm process?? 2 years ?? 4 years ??
  • eek2121 - Saturday, January 11, 2020 - link

    AMD could easily target 4 cores/8 threads. They would absolutely destroy Intel in frequencies and IPC alone. However, they chose to (rightly) differentiate themselves by going with 8cores with a similar clock speed.
  • RSAUser - Sunday, January 12, 2020 - link

    There are always use cases, and allowing for 40/50W boost usually results in a better user experience as most user tasks are short.

    Allowing for 28W constant is good for higher performance use cases, e.g. Gaming or something.

    This chip should probably also be pretty good silicon, so I'm guessing that its general draw will be quite low since they'll downlock dynamically (as pretty much all CPU nowadays do, ryzen has an advantage here from 2nd Gen onward).
  • ksec - Friday, January 10, 2020 - link

    I hope I am wrong. But I am going to guess they are prepping those 10nm 9W and 28W Stock for a certain product that are well known to have high volume in the market.

    MacBook Air ( 9W ) and MacBook Pro 14" ( 28W) .
  • tipoo - Friday, January 10, 2020 - link

    Hopefully. Though Ryzen 4000 putting 8 cores in the same wattage kind of takes the wind out of this long awaited update.
  • timecop1818 - Friday, January 10, 2020 - link

    Yeah but nobody cares about AMD's buggy slow power guzzling trash tho.
  • Korguz - Friday, January 10, 2020 - link

    so says the intel shill known as timecop1818
  • tipoo - Friday, January 10, 2020 - link

    We'll have to see how their battery life claims hold up, if they really did address it and come close to bridging the gap, 8 cores instead of 4 in 15-28W is hard to argue with.
  • Alistair - Friday, January 10, 2020 - link

    New Ryzen 4000 6 core runs at 2.1 Ghz at 15W, Intel only runs at 1.1 Ghz. AMD is almost twice as fast restricted to 15W. You're in for a surprise when Ryzen 4000 hits the market.
  • Korguz - Friday, January 10, 2020 - link

    hopefully...
  • timecop1818 - Friday, January 10, 2020 - link

    Yeah that's why every premium laptop vendor is just DYING to release AMD models huh?
  • Sahrin - Friday, January 10, 2020 - link

    I mean, AMD doubled their number of design wins in a single year...so something's going on for sure.
  • timecop1818 - Saturday, January 11, 2020 - link

    So.... from one to... two? Haha.
  • Spunjji - Saturday, January 11, 2020 - link

    @timecrap1818 - More than 100 this time. You commented extensively on the article where that was mentioned... Did you not read it? 🤔
  • timecop1818 - Saturday, January 11, 2020 - link

    No, I never read the articles. And I just celebrated AMD releasing garbage by buying a new i7-1065G7 laptop.
  • Korguz - Saturday, January 11, 2020 - link

    spoken like an intel shill timecop...
  • sarafino - Sunday, January 12, 2020 - link

    I hope it wasn't a Dell XPS. Sounds like everyone with an XPS and the 1065G7 are throttling all the way down to 1.7 GHz at 60c under load. Power limit exceeded even on 10nm? Eeek.
  • Spunjji - Monday, January 13, 2020 - link

    That candid admission from timecrap makes a lot of sense. Time to ignore everything they say :)
  • Retycint - Friday, January 10, 2020 - link

    Well, yes, actually
  • Alistair - Saturday, January 11, 2020 - link

    yes they are, we are talking about the upcoming Ryzen 4000 series CPUs and you'll see them everywhere for sure
  • Tigersam - Saturday, January 11, 2020 - link

    Thats max boost speed (Up to) as stated in the slides. It does not mean those 6 cores can run at 2.1GHz and will consume just 15W... No it does not work like that.

    What will happen is that when all 6 cores are active, the clockspeed and voltage will be constantly adjusted so it stays within that 15W limit. 2.1GHz is possible only very light loads. At full load, I am dead sure it will throttle much lower than that.

    Also 1.1GHz for the Intel CPU means base speed. Max is 3.3-3.6GHz depending on models. Same thing here. The CPU will adjust its speed to stay within power limits. Max possible speed for all 4 cores is around 3.6GHz. At full load, it will throttle down.

    Lastly, there is also thermal limits. If the CPU gets too hot under load, it will also throttle down. Laptops have limited cooling capacity compared to desktop.
  • SolarBear28 - Saturday, January 11, 2020 - link

    I assume we're talking about the Ryzen 5 4600U? Base clock 2.1 GHz, boost clock 4.0 GHz. 15 W TDP.
  • Alistair - Saturday, January 11, 2020 - link

    exactly, i was talking about base clocks, but he couldn't believe his eyes
  • Tigersam - Sunday, January 12, 2020 - link

    Wrong, boost clock of 4GHz is only possible with 1 core. Not all cores. 2.1GHz isn't the base speed. Its just the max with all cores running.
  • PixyMisa - Saturday, January 11, 2020 - link

    Base clock.
  • Alistair - Saturday, January 11, 2020 - link

    you got it exactly wrong: i quoted the BASE CLOCKS AT 15W, not the boost clocks, you're seemingly not aware of how AMD's mobile CPU will be almost twice as power efficient vs Intel

    not only was the desktop model already almost twice as efficient, that model had a 14nm IO die with 7nm chiplets

    the new mobile designs are monolithic 7nm, totally different CPUs, and VERY power efficient
  • Alistair - Saturday, January 11, 2020 - link

    https://cdn.videocardz.com/1/2020/01/AMD-Ryzen-400...

    and

    https://cdn.videocardz.com/1/2020/01/AMD-Ryzen-400...
  • Fataliity - Saturday, January 11, 2020 - link

    It will do an all-core around 2.9-3.4 GHz. EPYC cpu's can already do 3.0 at 2-3W TDP's. If they use the same quality cores on the 15w, and the high leakage on the 45w we should be fine. 2W x 8 core = 16W. 3Wx8Core = 24W
  • smilingcrow - Saturday, January 11, 2020 - link

    Is that 2-3W range just for the cores or does it also include the whole SoC?
  • Tigersam - Sunday, January 12, 2020 - link

    2-3W is idle.... not full load. Try running AVX and see if each core consumes 3W....
  • smilingcrow - Sunday, January 12, 2020 - link

    No way they are idling at 3W per core as that would be 64x3 = 192W just for the cores.
    Add the uncore to that and you are looking at over 250W for the whole chip.
  • Nicon0s - Monday, January 13, 2020 - link

    He said 2-3W/core at 3ghz.
    In idle most cores would be at 0% utilization so it should be lower than that.
    Face it Tigersam you were wrong.
  • Nicon0s - Monday, January 13, 2020 - link

    2.1GHz is the base clock, I hope you do understand what that means.
  • peevee - Wednesday, January 22, 2020 - link

    2GHz on AMD is base speed.
  • peevee - Wednesday, January 22, 2020 - link

    And better GPU too...
  • hansmuff - Friday, January 10, 2020 - link

    buggy slow power guzzling as in faster than Intel on the HEDT, less vulnerabilities, less power use for 16 cores than Intel's upcoming 10 core 300+W abortion of a chip?
  • yeeeeman - Saturday, January 11, 2020 - link

    Desktop and mobile platforms are entirely different things with different constraints and design considerations. The fact that AMD is better in HEDT market doesn't mean that automatically it will be the best in mobile.
  • eek2121 - Saturday, January 11, 2020 - link

    It surely does. AMD is just more efficient, period. Take a look at the Renoir slides. Ice Lake is DOA. I am not an AMD fanboy, but DAMN.
  • Nicon0s - Monday, January 13, 2020 - link

    You are correct, AMD's mobile design are even more efficient than their desktop ones. So they should fair even better in efficiency.
  • ksec - Saturday, January 11, 2020 - link

    Not this year. I still dont see how Apple could solve the Thunderbolt issue. The Spec is out, but royalty and certification is still not. So as long as Apple is held hostage by Intel with Thunderbolt, AMD on Mac is still wishful thinking.
  • tipoo - Saturday, January 11, 2020 - link

    USB4 will absorb the standard, no? Doesn't that mean AMD would get it with USB4 by default?

    There's Thunderbolt 4 announced, but it's not clear what it is really, from how many times faster than USB 3.2 Intel said, it sounds like the peak speed might not change.
  • ksec - Tuesday, January 14, 2020 - link

    That is the frustration. Thunderbolt 3 Spec is out, you can design it. But as long as the certification is still done by Intel, you are at their mercy. The same is with USB4, the Thunderbolt compatibility is an *optional* part.
  • eek2121 - Saturday, January 11, 2020 - link

    Several AMD boards have Thunderbolt.
  • TheinsanegamerN - Saturday, January 11, 2020 - link

    Which AMD boards have 4x thunderbolt ports capable of driving accessories like those available for apple products?

    Oh, right, none of them.
  • duvjones - Monday, January 13, 2020 - link

    You are not about to see AMD move on Thunderbolt till USB4 is prevalent. It is really the only way to work around the patents and trademarks, despite Intel giving Thunderbolt3 away to the USB forum.
  • Spunjji - Monday, January 13, 2020 - link

    There's no practical reason why they haven't done that, though - it's down to cost and not wanting to waste 4 PCIe lanes on seldom-used ports. They have to pay Intel for each controller, then spend extra to find the board space and metal layers in the PCB to route them - why do that for twice as many ports when a desktop will rarely (if ever) need more than 2 Thunderbolt ports?
  • sarafino - Tuesday, January 14, 2020 - link

    ASRock offers an X570 boards with TB3. Most don't waste PCIe lanes on so many TB ports.
  • Tigersam - Saturday, January 11, 2020 - link

    No no, cores don't mean anything in laptop CPUs. This is why.....

    All CPUs are limited by power (4800U has 15W tdp, can increase to 25W).

    Although 4800U has 8 cores and 16 threads, its still limited by power. If only 1 core is active, it can boost up to 4.2GHz (stated in the slides). If all 8 cores are active, how fast it can run depends on power. The slides say up to 1.8GHz. So, 8 cores at 1.8GHz max..... If all cores are running at 100%, it will likely throttle further to stay within that 15W TDP.

    Although I didn't read anything about power, its likely to have a short term and long term boost similar to Intel. IT may be possible to boost above the 15-25W limit for a short time (eg. 20-30s).

    So, more cores doesn't mean better performance due to this power limitation.
  • SolarBear28 - Saturday, January 11, 2020 - link

    More cores means a lot for multithreaded performance. It's more power efficient to run 8 cores slowly than it is to run 4 cores quickly. So you can get more done in the same TDP with more cores.
  • tipoo - Saturday, January 11, 2020 - link

    More cores in the same power limitations, with the core designs being near equivalent, almost always means more total aggregate performance, it takes less power to add a core than to scale frequency, the whole reason we went multicore.
  • eek2121 - Saturday, January 11, 2020 - link

    AMD CPUs typically do not throttle below base clock except in rare situations (prime95 small ffts is the only workload I have seen do it). The all core boost is likely around 3.4-3.6 GHz, with only extremely heavy workloads dropping to 1.8 GHz. If AMD indeed shrunk the IO die and moved to LPDDR4X, that gives each core the power budget it needs to maintain a high boost when needed.

    Furthermore, the boosting algorithm is fundamental to performance uplift: the smarter the chip is about boosting, the faster the chip performs.
  • oleyska - Sunday, January 12, 2020 - link

    I've found no workload pulling my amd cpu's below base. even with small ffts.
    Epyc(zen2) chips can drop 25-50 mhz below briefly with avx loads but in general they don't in real world.

    Yet to try mobile, excited to test out these new zen2 laptop chips.
  • Nicon0s - Monday, January 13, 2020 - link

    These new AMD APUs don't have an IO die, they are monolithic designs.
  • Nicon0s - Monday, January 13, 2020 - link

    No, the slides say 1.8GHz is the base clock speed. Learn to read the slides.
  • vFunct - Saturday, January 11, 2020 - link

    Really need the 45W version for MacBook Pro 16"
  • peevee - Wednesday, January 22, 2020 - link

    Why would anybody choose 4c@1GHz@9W when they can have 4c@1.2GHz @7W from the same old Intel?
  • SolarBear28 - Friday, January 10, 2020 - link

    So Intel basically lied about the initial specs of Comet Lake.
  • Sahrin - Friday, January 10, 2020 - link

    Somebody badly needs to do a piece on exactly what the fuck is wrong with the 10nm process, because not being able to validate basic IP blocks like an IMC is beyond just 'having trouble with yields.'
  • SolarBear28 - Friday, January 10, 2020 - link

    The thing is, Intel supports LPDDR4x with Ice Lake (10nm). But on their old 14nm process (Comet Lake) they don't (at least not until this new batch).

    Labeling both Ice Lake and Comet Lake as "10th Gen" was even more ridiculous than it first seemed.
  • peevee - Wednesday, January 22, 2020 - link

    Second that. Behind the scenes from anon sources would be highly appreciated.
    I bet it has a lot to do with the focus on "diversity" etc instead of brains.
  • repoman27 - Saturday, January 11, 2020 - link

    Yes, badly.

    First they decided to rename Amber Lake Y 4+2 "Comet Lake Y", even though it's really just a 7W version of Whiskey Lake 4+2 paired with an ancient 22nm Sunrise Point PCH. Then they didn't bother to actually ship it.

    Next they convinced the media that all Comet Lake U parts support LPDDR4X, yet they still haven't released any that do. In fact, the Core i7-10710U is the only Comet Lake part announced / launched thus far that even requires new silicon and can't just be produced from the previous generation WHL-U 4+2 V0 CPU die.

    Intel knew they were releasing two versions of Comet Lake U, and that v2 with LPDDR4X support wasn't going to be part of the initial Q3'19 launch. Leaked roadmaps from over a year ago had the LP4x version launching in Q2'20, which is probably accurate. I don't understand the need for all the subterfuge. And it's ridiculous to provide the press with slide decks that clearly contradict the specifications posted on ARK and the datasheets which were subsequently made available.
  • HurleyBird - Friday, January 10, 2020 - link

    "Those names are still used externally, however Intel has stated to us in conversation that they want to move away from that segmentation as a marketing message."

    No, that would make too much sense.
  • patel21 - Friday, January 10, 2020 - link

    I really wish a laptop manufacturer puts this 28W chip in a 45W laptop chasis with 70-90 Whr battery.

    It would really showcase the performance of this chip with enough headroom at boost clocks.
  • patel21 - Friday, January 10, 2020 - link

    I would like to add that, XPS 15 would be the best choice for this chip. Its is proven to be incapable of cooling a 45W chip with dGPU, so this chip fits it perfectly, as it doesn't need a dGPU.
  • brantron - Friday, January 10, 2020 - link

    Spoiler alert: previous XPS 13 was ~30 watts heat dissipation for 15w CPUs set to 25w, so just look there.

    In the past, the 28 watt laptop parts were genuinely higher power, with eDRAM, 2x GPU, and a tangible difference in turbo clocks.

    And then there's 28 watt Ice Lake, which is now likely to launch alongside a faster 15 watt Ice Lake i7 of the same stepping. DOA.
  • Retycint - Friday, January 10, 2020 - link

    The base XPS 15 with integrated GPU already fits that description. It may not be able to cool both a CPU and a discrete GPU but CPU only seems to be fine for the cooling
  • skavi - Friday, January 10, 2020 - link

    the MBP already handles it quite well.
    https://www.notebookcheck.net/Apple-MacBook-Pro-13...
  • wr3zzz - Friday, January 10, 2020 - link

    Can you do passive cooling at 9W? The old Y-series fanless were 5W.
  • tipoo - Friday, January 10, 2020 - link

    You could definitely do it passively, it just depends on how much you want to trade off size (larger passive heatsinks) vs just adding a fan. The Surface Pros sustained 15W passively just fine.
  • Retycint - Saturday, January 11, 2020 - link

    Don't see why not, since manufacturers are free to change the TDP limits. So a fanless design might limit the processor to 5-7W, for instance
  • PeachNCream - Saturday, January 11, 2020 - link

    TDP limits for fanless systems depend greatly on the design. Cooling placement, airflow, mass, skin temps, and so forth all play a big role in whether or not a given OEM will opt for a fanless system. Lately, with shedding thickness a priority, there is less cooling capacity to work with in order to achieve a passively cooled system. Lots of older, thicker notebooks that were running 35W Sandy Bridge chips could easily be converted to passive 9-15W designs using modern processors because they were comparably thick (although a lot of those designs were remarkably thin for their time period as well). It's really all in what an OEM will determine sells the best and costs the least.
  • repoman27 - Saturday, January 11, 2020 - link

    "Intel confirmed that the reason no LPDDR4-based Comet Lake devices have been launched is that there will be a new stepping of CPUs, labeled K1 instead of B0, that will enable support for LPDDR4."

    CML-U thus far has been built from the A0 6+2 die and previous generation WHL-U 4+2 V0 die. The most recent B0 stepping was CFL-S 4+2. Not that K1 is wrong, just that B0 doesn't sound entirely right to me.
  • alufan - Saturday, January 11, 2020 - link

    Intel still playing the public for fools and frankly most of them believe it however there is a shift in perception coming AMD has started to get its act together on mobile as well but i believe Intel will still have the top end once all this plays out, next gen Ryzen though will redress the balance am sure
  • Fataliity - Saturday, January 11, 2020 - link

    if they really wanted to make a really fat laptop, they could take a 3950x in eco mode at 65W and cool it, and it will perform great. I guess they're signaling its not a big enough market to make it
  • jospoortvliet - Saturday, January 11, 2020 - link

    The idle draw would make it useless. Mobile platforms are quite different from desktop...
  • anactoraaron - Saturday, January 11, 2020 - link

    You would be talking about the Clevo NH57ADS. Support for 65w desktop ryzen 3000, shown at CES.

    Honestly this notebook is the most excited I've been in some time about a notebook release.
  • Fataliity - Saturday, January 11, 2020 - link

    Definitely. Sure the battery won't matter even if its giant, but everyone I know these days games on a laptop just because its more convenient.
    Desktop is basically dead. It's all MSI / Alienware high-end 2k laptops. And from experience the Alienware is junk (keyboard and screen) seen 2 of them fail in a year. MSI is going 2 years later still no issues. Also gets better performance and better parts for a cheaper price. MSI was around 2k, Alienware was around 2800.
  • Fataliity - Saturday, January 11, 2020 - link

    But the Clevo does sound awesome. They should make either a thin and light super thin power-efficient laptop, or a nice big work / gaming laptop. Who really cares much about the middle? You can just up the TDP from bios from 15 to 25 as long as the design can handle the TDP.
  • danielfranklin - Sunday, January 12, 2020 - link

    You mean the 28w Apple chip will ship when Apple update their 28w laptop, shocking...
  • canukstorm - Monday, January 13, 2020 - link

    Why call it the Apple chip? I would presume other PC vendors would utilize it as well.
  • Spunjji - Monday, January 13, 2020 - link

    Historically, they don't - most PC vendors either go all-out on thin-and-light with a 15W CPU, middle-ground with 15W CPU and MX / Max-Q graphics, or put it all down on power with 45W CPUs and dGPUs. Apple are basically the only vendor who reliably release a device with the 28W CPU line, and the first models in it were designed at their request (they needed more iGPU performance for the Retina Macbook designs).
  • canukstorm - Monday, January 13, 2020 - link

    Ok, I see. Thanks.
  • Spunjji - Monday, January 13, 2020 - link

    Apple can't update their 28W laptop until Intel ship the CPUs. You seem to have cause/effect the wrong way around here.
  • SharonTTurner - Monday, January 13, 2020 - link

    If the 15W chip can TDPup to 25W, can the 28W chip do something similar?
  • canukstorm - Monday, January 13, 2020 - link

    So we should see the next MBP using these processors around Fall 2020 then.
  • peevee - Wednesday, January 22, 2020 - link

    Intel's 10nm is still a disaster.
    Compare 9W 4 cores at 1GHz vs their own 14nm 7W (!!!) 4 cores at 1.2GHz.

Log in

Don't have an account? Sign up now