09:47PM EDT - When the big trade shows roll around, this is the time for the big companies in our sphere to announce their next biggest hardware, or update us on what is to some. AMD had some really big launches at the top of the year, with Ryzen 5000 for desktop and mobile, Radeon for desktop, and then a bit later we saw EPYC on Zen 3 come to market. This year at Computex, CEO Dr. Lisa Su heads up AMD's keynote presentation, and we're here ready to live blog the announcements as they come in. Join us at 10pm ET (10am Taiwan local time)!

09:55PM EDT - Four minutes to go

09:58PM EDT - Here's the AMD Youtube stream, for anyone who wants to watch along with the live blog: https://www.youtube.com/watch?v=gqAYMx34euU

10:00PM EDT - Two minute countdown

10:02PM EDT - Intro video

10:03PM EDT - Share new products

10:03PM EDT - Starting with the pandemic

10:03PM EDT - Positive impact with vaccination

10:04PM EDT - Crowd in attendance, AMD employees

10:04PM EDT - High performance computing

10:04PM EDT - We are in a megacycle

10:04PM EDT - transformation of devices, ai

10:04PM EDT - Push the envelope in High Perf Computing

10:05PM EDT - Preview of advanced technology at the end

10:05PM EDT - March was 3rd Gen AMD EPYC Milan, based on Zen 3

10:05PM EDT - Deep partnerships across the server ecosystem

10:05PM EDT - 100+ Epyc server platforms, 400+ Epyc instances by end of year

10:05PM EDT - AMD-powered cloud services

10:06PM EDT - Microsoft 365, teams, twitter, zoom, tencent meeting

10:06PM EDT - Momentum with 3rd Gen EPYC

10:06PM EDT - 220+ world records on Milan

10:06PM EDT - WR in Int, FP, Java, DB, Analytics

10:07PM EDT - Doubled solutions compared to previous gen

10:07PM EDT - HCI, converged infrastructure, HPC, analytics

10:07PM EDT - Since launch, compare against Ice Lake

10:07PM EDT - e-commerce perf

10:07PM EDT - This is java

10:07PM EDT - With SLA

10:07PM EDT - 2P 7763 vs 2P 8380

10:08PM EDT - POS requests, online purchases, analytics

10:08PM EDT - 200k vs 300k perf

10:08PM EDT - No mention of Arm / Ampere Altra

10:09PM EDT - 50% better TCO on this workload

10:10PM EDT - Now the new architectures

10:11PM EDT - New AMD APUs into the market

10:11PM EDT - 5700G and 5600G

10:11PM EDT - 8-core and 6-core

10:11PM EDT - 65W

10:12PM EDT - Performs above the competition

10:12PM EDT - Both models available August 5th

10:12PM EDT - $359 and $259, worldwide launch

10:12PM EDT - Now RDNA2

10:13PM EDT - First cards in Q4 with 6900XT, 6800XT, and 6800

10:13PM EDT - 6700 was early this year

10:13PM EDT - RDNA2 in the consoles

10:13PM EDT - Just getting started with RDNA2

10:13PM EDT - RDNA2 coming to Tesla!

10:14PM EDT - Embedded ryzen APU in the infotainment

10:14PM EDT - discrete RDNA2 that kicks in 10 Teraflops when AAA is needed

10:14PM EDT - Tesla is the next platform for triple A gaming

10:14PM EDT - High perf mobile RDNA2

10:14PM EDT - Custom Graphics IP to Samsung with Ray Tracing and Variable Rate Shading

10:15PM EDT - Samsung to provide details later this year

10:15PM EDT - Now notebooks

10:15PM EDT - Scott Herkelman to the stage

10:15PM EDT - Now that's a smug face

10:16PM EDT - +27% YoY gaming laptop growth

10:16PM EDT - Looking at Gen on Gen perf

10:17PM EDT - Up to 50% more perf at the same power

10:17PM EDT - 43% lower power at same perf

10:17PM EDT - +77% more perf Gen on Gen off the battery

10:17PM EDT - *on the battery

10:17PM EDT - announcing RX 6000M series

10:18PM EDT - 6800M - 2300 MHz, 12GB GDDR6

10:18PM EDT - ROG Strix 15

10:19PM EDT - 120 FPS on RE:Village at 1440p Max

10:19PM EDT - gaming on battery leadership

10:19PM EDT - AAA and eSports

10:20PM EDT - Ultimate gaming on the go experience, best in class battery perfomance

10:20PM EDT - Radeon Chill for longer gaming

10:20PM EDT - Also two other products

10:20PM EDT - RX 6700M

10:20PM EDT - 100FPS at 1440p max

10:21PM EDT - RX 6600M

10:21PM EDT - HP Omen 16 shipping later this month

10:21PM EDT - AMD FidelityFX Super Sampling

10:22PM EDT - 10s of millions of gamers on laptops need the best experiences

10:22PM EDT - high performance upscaling required

10:22PM EDT - AMD FSX for all gamers

10:22PM EDT - *FSR

10:22PM EDT - spatial upscaling

10:22PM EDT - FSR enables broad adoption

10:23PM EDT - 4 different quality settings

10:23PM EDT - Supported on all Vega or higher GPUs

10:24PM EDT - Available for the entire industry to use on GPUOpen

10:24PM EDT - Works on NVIDIA hardware as well

10:24PM EDT - works on all 10-series NVIDIA cards and newer

10:25PM EDT - Support in 10 game engines already

10:25PM EDT - support on 100 GPUs and CPUs and competitor products

10:25PM EDT - First games support with FSR on June 22nd

10:25PM EDT - Frank Azor to the stage

10:27PM EDT - Laptops have no room for error

10:27PM EDT - A great gaming laptop isn't easy

10:27PM EDT - It goes beyond CPU+GPU

10:27PM EDT - power, cooling, display, storage, IO, thermals, size

10:28PM EDT - AMD Advantage

10:29PM EDT - AMD Advantage laptop buyers know it was designed for a great gaming experience

10:29PM EDT - Requires AMD Ryzen 5000 CPU, RX6000M graphics, Radeon Software

10:29PM EDT - Exclusive AMD smart technologies

10:29PM EDT - AMD Smart Access Memory

10:30PM EDT - SAM comes to advantage with RX 6000M

10:30PM EDT - Also AMD Smartshift

10:30PM EDT - shifts power between CPU/grapics

10:30PM EDT - +11% gaming performance

10:31PM EDT - All AMD Advantage laptops have 144 Hz+ displays, 300 Nits+, IPS/OLED

10:31PM EDT - All about the gamers

10:32PM EDT - All AMD Advantage will have at least one NVMe x4 drive

10:32PM EDT - low temps on the WASD and arrow keys

10:32PM EDT - First AMD A+A is ROG Strix G15

10:33PM EDT - 5900HX, 300 Hz display, RX 6800M

10:33PM EDT - Select countries in June

10:33PM EDT - HP Omen, 16-inch 5900HX + 6600M, 165 Hz

10:33PM EDT - 1080p gaming

10:34PM EDT - More models coming soon

10:35PM EDT - FSR available on GPUOpen on MIT license

10:35PM EDT - Now advanced technology

10:35PM EDT - X3D packaging

10:36PM EDT - roadmap to 5nm with first Zen 4 products next year

10:36PM EDT - TSMC 3d fabric

10:37PM EDT - First application is 3D vertical cache

10:37PM EDT - stacking 64 MB on top of a zen 3 chiplet

10:37PM EDT - TSVs

10:37PM EDT - 2 TB/sec bandwidth

10:38PM EDT - 192 MB L3 in a single package

10:38PM EDT - 200x interconnect density

10:39PM EDT - Denser connection with IP

10:39PM EDT - No solder bumps, TSVs only

10:39PM EDT - 3D V-Cache technology

10:39PM EDT - 184 fps vs 206 FPS

10:40PM EDT - 4 GHz clock speed fixed

10:40PM EDT - +12 %

10:40PM EDT - 12 core / 24 thread

10:40PM EDT - +15% faster gaming on average

10:40PM EDT - One entire architecture gain just from vcache

10:41PM EDT - Starting production end of the year

10:42PM EDT - That's a wrap. That was unexpected.

Comments Locked

32 Comments

View All Comments

  • Chaitanya - Monday, May 31, 2021 - link

    Good to see AMD keeping FSR platform agnostic.
  • Smell This - Tuesday, June 1, 2021 - link


    Doc Su and her team has "roped the dopes," again. Nicely done, AMD ... it's not just FidelityFX, though the cross-platform open approach is a winner all-around.

    AMD is taking on the industry's best __ with each industry's best stuff __ and raising the performance and efficiency bar, incrementally and on-map. Whether HPC, mobile, enterprise, desktop or gaming. Good for them and good for all of us.

    Competition is a good thing. Can't wait to hear more about V-Cache, 'TSMC 3d fabric,' and chiplet (GloFo 12nm?) design.
  • Byte - Tuesday, June 1, 2021 - link

    Probably genius FidFX works on both. Game devs will probably just focus on that as they can optimize on the consoles and port easier to PC with that one tech. AMD can hardware accelerate it and might perform worst on Nvidia showing them leading in many benchmarks.
  • flashmozzg - Tuesday, June 1, 2021 - link

    It worked for FreeSync, they hope it'll work for FSR.
  • nandnandnand - Monday, May 31, 2021 - link

    Zen 3+ = mostly the same as Zen 3 but with 3D V-Cache?
  • Makaveli - Monday, May 31, 2021 - link

    That does make sense.
  • SaturnusDK - Monday, May 31, 2021 - link

    Most likely.

    Consider the implications when used with 8 CCDs on a Zen3+ based EPYC. 1536MB of L3 cache!
  • Yournotme - Monday, May 31, 2021 - link

    It would be 768MB for 8 chiplets. The 192MB was for 2 chiplets.
  • SaturnusDK - Monday, May 31, 2021 - link

    I was talking about a 2P platform though.
  • Kamen Rider Blade - Monday, May 31, 2021 - link

    So 96 MiB for 8x Cores then?

    That boils down to 12 MiB of L3$ per Core?

Log in

Don't have an account? Sign up now