AMD Reveals 16-Core, 4.7 GHz Boost Ryzen 9 3950X: $749, September Arrival

AMD CEO Lisa Su recently announced the company's new Ryzen 3000 series lineup at Computex 2019, but left out one official announcement for its Next Horizon Gaming Tech Day here in Los Angeles: The 16-core 32-thread Ryzen 9 3950X for $749. This new processor will land on shelves in September. We have more coverage of the Zen 2 microarchitecture and X570 chipset here.

Paul Alcorn
Editor-in-Chief

Paul Alcorn is the Editor-in-Chief for Tom's Hardware US. He also writes news and reviews on CPUs, storage, and enterprise hardware.

  • Giroro
    I was hoping for $699 on the Ryzen 9 3950X, which is closer to (but still above) the ~$42/core of the 3900X, 3700X, and 3600X.
    At $750 the 3950X is ~$47/core
    But at least it's not the $800 it would have been had AMD chosen the $50/core pricing of the 3800X (which is a little overpriced compared to the rest of lineup). The 3800Xis in the middle of the stack and doesn't even have the prestige of being the world's first at anything.
    If the 3600 overclocks well, then it's looking like it will be an outstanding value at ~$33/core
    Reply
  • JQB45
    I'm not ok with the 3600 clock speeds when comparing to AMD's own existing 1600 and 2600 parts, you're not getting much other then IPC unless this thing overclocks well. Personally I'll go to AMD's new mid-range during the holiday season and get a 3800 or 3800X. I don't need the 3900X or 3950X at this point.

    I am curious about the 3000-3500X line up and what that will bring to the truly budget constrained.
    Reply
  • jimmysmitty
    The biggest issue with more cores is software adoption. Everyone clamored that consoles using 8 core AMD CPUs would force game developers to utilize more cores. 6 years later and 8 cores doesn't mean squat in gaming, frequency still does as does IPC. So this will, per Lisa, be near Intel in IPC which means frequency will still win out the end game and in the mainstream market Intels 9th gen still holds that advantage pretty well from a stock perspective and even overclocking perspective.

    While I would say it would be a good workstation CPU I feel that the dual channel memory will be the biggest bottleneck for it as workstation applications not only like multiple cores it also likes a lot of fast memory. Dual channel will not allow this CPU to really stretch its legs. Maybe when we move to DDR5, which is rumored to double bandwidth at the same speed, it would be a good fit.

    With a new process and enhanced uArch we can hope that it will OC higher than Zen/Zen+. I think the biggest hold back was the 14nm process for Zen/Zen+, it was originally developed for low power products, while TSMCs 7nm should handle power better so maybe we will see good OCing out of it.

    I will wait till TH and others get their hands on this to make full judgements. My inital view is it will push Intel in some areas but not enough to make them finally start a good price war.
    Reply
  • mihen
    SOLD
    Reply
  • NightHawkRMX
    Games seem to be starting to use more cores over the past 2 years.
    4c i5 7600k beat the 6 core ryzen 5 in nearly all games 2 years ago.
    Now in new games the 1600 wins in nearly all games. Framerate on the 7600k is unstable but the 1600 is smooth.
    Reply
  • TCA_ChinChin
    jimmysmitty said:
    While I would say it would be a good workstation CPU I feel that the dual channel memory will be the biggest bottleneck for it as workstation applications not only like multiple cores it also likes a lot of fast memory. Dual channel will not allow this CPU to really stretch its legs. Maybe when we move to DDR5, which is rumored to double bandwidth at the same speed, it would be a good fit.

    I will wait till TH and others get their hands on this to make full judgements. My inital view is it will push Intel in some areas but not enough to make them finally start a good price war.

    Although AMD did emphasize multicore workstation benchmarks, I'm pretty sure they are marketing this more towards gaming since they literally said that the 3950x was the worlds first 16-core GAMING cpu. Even if people might think and even use the r9 for workstation tasks, AMD themselves are not heavily leaning towards it in their own marketing. They have the next generation of Threadripper which will be more appropriate and unfortunately probably more expensive.

    Intel either won't respond and lose significant market share or drop prices. This is good enough to the point where no matter how you spin it, these AMD products are superior to Intel's current lineup. If they don't change things up at Intel and rest on their war chest from sandbagging to keep up the mirage of superiority, they deserve to lose.
    Reply
  • tntom
    Any word on the Threadripper CPUs? Some have speculated they may have served their time but will be dropped. 16C/32T used to be TR territory. Where would TR top out? 32/64 perhaps with 64 PCIe lanes like Zen+? or could they take it up to 64/128?
    Reply
  • AllanGH
    I'll definitely give it a looksee after the platform, as a whole, has had about a year to mature....but it did make me take a very interested look.
    Reply
  • TJ Hooker
    I'd be curious to see once a 16C/32T Zen 2 thread ripper comes out how much of a difference dual vs quad channel memory will make.
    Reply
  • Rdslw
    tntom said:
    Any word on the Threadripper CPUs? Some have speculated they may have served their time but will be dropped. 16C/32T used to be TR territory. Where would TR top out? 32/64 perhaps with 64 PCIe lanes like Zen+? or could they take it up to 64/128?
    I think they will make 16/32 + 24/48 + 32/64 MAYBE 48/96 + 64/128 TR's with 64 or 80 PCIE lines.
    (4 for sure 8 -> doubt chiplets)
    it was useful af for content creators and as its not high volume thing, still, a lot of would buy.

    they have cheap cores AND it scales better in bigger cpu's.
    expect TR to rule in 6-12 months.
    Epycs will go for 8cores x 8chiplets as lowest and imho an option for 8 cores x 16 chips will appear soon.
    Reply