New Upcoming ATI/AMD GPU's Thread: Leaks, Hopes & Aftermarket GPU's

Discussion in 'Videocards - AMD Radeon' started by OnnA, Jul 9, 2016.

  1. Goiur

    Goiur Maha Guru

    Messages:
    1,079
    Likes Received:
    372
    GPU:
    ASUS TUF RTX 3080
    If its out q2 of 2019, a 2080 level perf even at 400-500€ its a gamer changer. 2080 its around 700-900€ right now, a price drop should be in order if AMD can hit that mark.
     
    Maddness likes this.
  2. OnnA

    OnnA Ancient Guru

    Messages:
    12,370
    Likes Received:
    3,066
    GPU:
    Vega XTX LiQuiD
    So my Leak was True?
    We will have Vega 2 7nm or VII nm ;)
    Now we need to do the math, how fast will it be?

    @JonasBeckman ? Bring the Crystal ball :p
    Faster HBM2 4x4GB, Better GPU clocks & maby some enchancements.

    Note:
    Vega XTX at ~1757MHz/1200HBM2 gives >28k GPU pts. in 3Dmark

    Drv. 18.12.1.1 (Our Bratan' from overclock.net)
    -> https://www.3dmark.com/fs/17370977

    Note no2:
    So if Vega2 7nm will be 4x4 HBM2 (~1TB/s) and >1800MHz then Yes 29k-30k is possible
    Im talking 7nm "shrink" of My Vega, but if they do some magic on silicon level then that score can be even at 32k ballpark.

    [​IMG]

    [​IMG]
     
    Last edited: Dec 12, 2018
    cryohellinc likes this.
  3. OnnA

    OnnA Ancient Guru

    Messages:
    12,370
    Likes Received:
    3,066
    GPU:
    Vega XTX LiQuiD
    SAPPHIRE To Release Black Shroud Nitro+ Radeon RX 590

    [​IMG]
     
  4. cryohellinc

    cryohellinc Ancient Guru

    Messages:
    3,352
    Likes Received:
    2,674
    GPU:
    RX 5700 XT/GTX 1060
    I love the minimalistic approach to Vega 2 logo.
     
    OnnA likes this.

  5. OnnA

    OnnA Ancient Guru

    Messages:
    12,370
    Likes Received:
    3,066
    GPU:
    Vega XTX LiQuiD
    AMD/ATI Radeon Software Adrenalin 2019 Edition

    As has become tradition, AMD is yet again on the cusp of launching its next big driver package as a holiday gift to all of its Radeon fans.
    Adrenalin 2019 Edition is the fifth consecutive installment in AMD’s big yearly GPU driver package releases since it all began back in 2014 with the Omega Driver package.

    Performance

    The only slide with meaningful performance figures compares the new driver to 17.12.1 driver.
    Most of the games in this comparison were not even out when 17.12.1 was available, performance increased (up to) 15% for Radeon RX 570.

    Radeon Advisors
    • Game Advisor – “provides game settings guidance for a personalized & improved experience”
      This tool works in conjunction with Radeon Overlay. What it does is rather simple: it suggests how to improve performance and frame rating by adjusting game settings.
    • Settings Advisor – this is designed for new users who never used Radeon Settings before. It recommends which features should be enabled (VSR, FreeSync etc.) based on measured performance.
    • Upgrade Advisor – system analyzer for “minimum and recommended game compatibility”
    Wattman
    • Auto-GPU & Memory Overclocking
    • Auto-GPU undervolting
    • Temperature-dependent fan curves
    • Unlocked DPM states for RX Vega
    • Memory tuning
    Display Technologies
    • FreeSync 2 HDR – improved auto-tone mapping for a more detailed experience
    • Virtual Super Resolution – support for 21:9 displays
    Radeon Overlay
    • Display Settings – in-game controls: enhanced sync, FreeSync, per-game settings
    • Wattman – in-game adjustment for: gpu frequency, gpu voltage, gpu temperature, memory timing, memory frequency, load/save profiles
    • Refined performance metrics – framerate and frametimes – adjust colors, columns, position, size and transparency
    AMD Link
    • QR Code linking
    • Voice Control – streaming, recording, screenshots, instant replay, min/avg/max fps, gpu temp, gpu clocks, mem clocks, fan speed
    • Wattman – same settings as in Radeon Overlay
    • Enhanced Performance Metrics — detailed analytics (basically turning a phone into a benchmarking tool).
    • ReLive – View, Edit, Stream recordings/screenshots from a mobile device
    AMD ReLive
    • In-Game Replay
    • Scene Editor
    • GIF Support
    AMD ReLive streaming
    • Wireless streaming up to 4K60FPS to mobile devices
    • Free on Android and IOS
    • 70ms (AMD) vs 125ms (competitor) responsiveness
    • Works with AMD Link (through Game Explorer tab)
    • AMD ReLive for VR: support for HTC Vive Focus, Oculus GO, Samsung Gear VR, Google Daydream
      — features: 3rd party Bluetooth compatibility, up to 1440x1440p
     
    Last edited: Dec 11, 2018
    Maddness likes this.
  6. OnnA

    OnnA Ancient Guru

    Messages:
    12,370
    Likes Received:
    3,066
    GPU:
    Vega XTX LiQuiD
    Image Quality comparison IQ

     
  7. OnnA

    OnnA Ancient Guru

    Messages:
    12,370
    Likes Received:
    3,066
    GPU:
    Vega XTX LiQuiD
    JEDEC Updates HBM Standard With Up To 24 GB Memory & 307 GB/s Bandwidth Support Per Stack

    JEDEC has published an update for the high-bandwidth memory standard which allows higher capacities and faster pin speeds.
    With the update, system manufacturers adopting HBM can leverage from higher memory capacities and various device configurations as HBM memory serves several applications include Graphics, High-Performance Computing, Server, Networking and Client applications.

    JEDEC Published New Update For High-Bandwidth Memory Standard – Higher Memory Capacities of Up To 24 GB, Faster Pin Speeds

    The existing HBM2 memory allows for up to 8 GB memory in 4-Hi or 8-Hi stacks.
    The products that make the most use of the HBM2 standard currently in the graphics industry are the Tesla V100 that is produced by NVIDIA and the upcoming Instinct MI60 from AMD.
    Both cards would feature 32 GB of HBM2 memory and feature a Terabyte of bandwidth (900 GB/s for the Tesla V100).
    The key here is that these cards make use of 4 HBM2 stacks, each with 8 GB of memory per stack.
    These can be either 4-Hi or 8-Hi, depending on the memory configuration (Tesla V100 is also available in 16 GB HBM2 variant).

    ===
    JEDEC standard JESD235B for HBM leverages Wide I/O and TSV technologies to support densities up to 24 GB per device at speeds up to 307 GB/s.
    This bandwidth is delivered across a 1024-bit wide device interface that is divided into 8 independent channels on each DRAM stack.
    The standard can support 2-high, 4-high, 8-high, and 12-high TSV stacks of DRAM at full bandwidth to allow systems flexibility on capacity requirements from 1 GB – 24 GB per stack.


    This update extends the per pin bandwidth to 2.4 Gbps, adds a new footprint option to accommodate the 16 Gb-layer and 12-high configurations for higher density components, and updates the MISR polynomial options for these new configurations.
    Additional clarifications are provided throughout the document to address test features and compatibility across generations of HBM components.


    via JEDEC

    ====
    With the updated JEDEC standard, HBM will be updated to support 24 GB of memory per stack. The memory would be made up of 16 Gb dies and can scale all the way from 1 GB to 24 GB per stack depending on the height of the stacks themselves.
    The standard can support 2-Hi, 4-Hi, 8-Hi and 12-Hi TSV stacks and each would feature a 1024-bit wide bus interface.

    A single stack thus will offer speeds of 307 GB/s. If you look at the topmost configuration that one can get out of the new standard, you get 96 GB of memory and 1.2 TB/s bandwidth along a 4096-bit wide interface.
    This is an unprecedented amount of memory and bandwidth you can get from a single chip!

    [​IMG]

    THX to WCCFtech
     
    Last edited: Dec 18, 2018
    Deleted member 213629 likes this.
  8. Maddness

    Maddness Maha Guru

    Messages:
    1,456
    Likes Received:
    672
    GPU:
    3080 Aorus Xtreme
    With how fast GDDR6 memory is, it will be interesting to see if AMD sticks with HBM moving forward.
     
  9. Elder III

    Elder III Ancient Guru

    Messages:
    3,714
    Likes Received:
    320
    GPU:
    Both Red and Green
    Have there been any more leaks or rumors about the performance (and release date) of AMD's next gen series? I'm going to likely be in the market for one, but I need to decide what to do for the near future if they aren't releasing until Summer 2019.
     
  10. Maddness

    Maddness Maha Guru

    Messages:
    1,456
    Likes Received:
    672
    GPU:
    3080 Aorus Xtreme
    Nothing more unfortunately. Hopefully AMD will spill the beans at CES in January.
     

  11. OnnA

    OnnA Ancient Guru

    Messages:
    12,370
    Likes Received:
    3,066
    GPU:
    Vega XTX LiQuiD
    AMD Will Be Launching The Ryzen 3000 Series CPUs, APUs And A Radeon GPU At CES

    I have received confirmation that AMD will be launching 3 classes of product at their CES keynote: the Ryzen 3000 Series CPUs, Ryzen 3000 Series (APUs?) with Vega Graphics and a consumer Radeon graphics card – which is probably the Radeon Vega II.
    Lisa will also be talking in depth about AMD’s 7nm progress at the CES keynote.

    Update: AMD to launch a Radeon (consumer) GPU, Ryzen 3000 series and discuss 7nm progress at CES 2019 Keynote

    Now the Ryzen 3000 series is fairly straightforward. We have been hearing about the upcoming Picasso series for quite a while now and there is even talk on the grapevine that this will be featured in an upcoming XBOX console.
    That said, AMD will be launching both the CPU and APU SKUs at CES 2019 at their Keynote. The APUs will have vega graphics and will be competing against Intel/NVIDIA combos in the mobility segment.

    Perhaps the more interesting news is the confirmation that AMD will be launching a Radeon, consumer graphics cards. This much I have confirmed. As for what it is, based on my past talks with some industry insiders,
    I believe this is going to be the Radeon Vega II – the 7nm consumer variant of the Vega 20 GPU. This is something that I was told was one of the last projects of Mike Rayfield and should be the one AMD launches at CES 2019.
    (Update 4:39 AM 12/25/2018 PKT) Availability will be in 1H 2019 after the CES launch.

    That’s it for the update if you want to stick to our source and verified data. If you want to read on some more juicy (but unvetted) deets, here they are from past rumors and leaks:

    Rumored specifications and details based on past leaks:

    The Ryzen 3000U series has already had a couple of leaks and will likely be one of the lineups that AMD reveals at the show in January.
    The U stands for Ultra Low Voltage variants and will be used in netbooks and laptops with battery in mind (ie not gaming laptops).

    Here’s the thing though, past rumors indicated that the 3000U series will be based on the 12nm process and not the 7nm one while as some others have indicated that the Ryzen 3000 CPU series will be based on the 7nm node so it is unclear which process the 3000 series will hail from.
    The 7nm 3000 series is supposed to feature Zen 2 and the new X570 PCH (while staying on the AM4) socket and will be a considerable performance update over the last generation.

    [​IMG]
     
    Maddness likes this.
  12. Maddness

    Maddness Maha Guru

    Messages:
    1,456
    Likes Received:
    672
    GPU:
    3080 Aorus Xtreme
    Sounds awesome. I still have my doubts about 7nm Vega consumer card though. I would rather it be 7nm Navi 10 GPU.

    Really really looking forward to RyZen 2 CPU's. I've been holding out for an upgrade and this will be it.
     
  13. bubz

    bubz New Member

    Messages:
    9
    Likes Received:
    0
    GPU:
    gtx 1070
    Will the new Radeons on 7nm be on the level of top Nvidia cards?
     
  14. Maddness

    Maddness Maha Guru

    Messages:
    1,456
    Likes Received:
    672
    GPU:
    3080 Aorus Xtreme
    The early ones, very unlikely. There is a second Navi due in 2020 that should be up there somewhere.
     
  15. bubz

    bubz New Member

    Messages:
    9
    Likes Received:
    0
    GPU:
    gtx 1070
    Oh in 2020 AMD will be like 4 years behind Nvidia :p
     

  16. Maddness

    Maddness Maha Guru

    Messages:
    1,456
    Likes Received:
    672
    GPU:
    3080 Aorus Xtreme
    Yeah but, they are both a aiming at different markets. Nvidia is focused on the high end, enthusiast markets. AMD is focused on mid range and entry level.
     
  17. bubz

    bubz New Member

    Messages:
    9
    Likes Received:
    0
    GPU:
    gtx 1070
    They are also making components for playstation 5 ?
     
  18. OnnA

    OnnA Ancient Guru

    Messages:
    12,370
    Likes Received:
    3,066
    GPU:
    Vega XTX LiQuiD
     
    MerolaC likes this.
  19. OnnA

    OnnA Ancient Guru

    Messages:
    12,370
    Likes Received:
    3,066
    GPU:
    Vega XTX LiQuiD
    Maddness likes this.
  20. PrEzi

    PrEzi Master Guru

    Messages:
    597
    Likes Received:
    444
    GPU:
    AMD 6900XT
    "AMD claims Navi will compete with Nvidia’s high-end… is it a more powerful GPU than we thought?"
    https://www.pcgamesn.com/amd-navi-high-end-more-powerful-gpu-than-expected

    [...]I asked if that meant his goal for Navi was to get back to the days where now-AMD would be fighting it out with Nvidia for the top-GPU bragging rights, and he absolutely agreed (David Wang). Which would be quite the thing, and somewhat counter to what we thought we knew about where the Navi GPUs are going to be positioned.[...]

    Now THAT would be awesome!
     
    OnnA likes this.

Share This Page