This site may earn affiliate commissions from the links on this page. Terms of use.

AMD-vs-NV

Yesterday, Nvidia took the wraps off its loftier-end GP100 GPU and gave usa a await at what its elevation-end HPC configuration would look like come Q1 2017. While this new carte is explicitly aimed at the scientific computing marketplace and Nvidia has said nothing about futurity consumer products, the information the company revealed confirms some of what we've privately heard about next-generation GPUs from both AMD and Nvidia.

If yous're thinking nearly using some of your taxation rebate on a new GPU or only eyeing the market in general, we'd recommend waiting at least a few more months before pulling the trigger. Information technology may even be worth waiting until the end of the year based on what nosotros now know is coming downwardly the pipe.

What to expect when you're expecting (a new GPU)

Showtime, a bit of review: We already know AMD is launching a new set of GPUs this summer, codenamed Polaris 10 and Polaris 11. These cores are expected to target the sugariness spot of the add-in-board (AIB) market, which typically means the $199 – $299 price segment. High-end cards like the GTX 980 Ti and Fury X may control headlines, simply both AMD and Nvidia ship far more than GTX 960s and Radeon R7 370s than they do top-end cards.

AMDGPU

Polaris 10 and xi are expected to use GDDR5 rather than HBM (I've seen the rumors that merits some Polaris SKUs might use HBM1 — it's technically possible, merely I think it exceedingly unlikely) and AMD has said these new GPUs will improve functioning-per-watt past 2.5x compared with their predecessors. The visitor's side by side-generation Vega GPU family, which arrives late this yr, is rumored to be the first basis-upward new architecture since GCN debuted in 2012 with four,096 shader cores and HBM2 memory.

We don't know even so what Nvidia'south plans are for any consumer-oriented Pascal cards, only the speeds and core counts on GP100 tell us rather a lot about the benefits of 16nm FinFET and how it will bear on Nvidia's product lines this generation.

Pascal-Chart

With GP100, Nvidia increased its core count past 17% while simultaneously ramping up the base of operations clock past xl%. Baseline TDP for this GPU, meanwhile, increased by xx%, to 300W. The relationship between clock speed, voltage, and power consumption is non linear, simply the GTX Titan Ten shipped with a base clock of 1GHz, only slightly higher than the Tesla M40'due south 948MHz. The GP100 has upwardly to 60 SM units (just 56 are enabled), which puts the total number of cores on-die at iii,840. That's 25% more cores than the sometime M40, but the die is just 3% larger.

We may not know details, but the implications are straightforward: Nvidia should be able to deliver a high-end consumer card with 30-40% higher clocks and significantly college core counts within the same price envelopes that Maxwell occupies today. We don't know when Team Green will get-go refreshing its hardware, only it'll nigh certainly be inside the next nine months.

Hither'due south the bottom line: AMD is going to start refreshing its midrange cards this summer, and it'd be unusual if Nvidia didn't have fresh GPUs of its own to encounter them. Both companies will likely follow with high-cease refreshes towards the end of the year or very early adjacent year, once more, probably inside short order of each other.

When waiting makes sense

There'southward a platitude in the tech industry that claims it'southward foolish to endeavour and fourth dimension your upgrades because engineering is always advancing. 10-12 years ago, when AMD and Nvidia were nearly doubling their tiptop-end operation every single twelvemonth, this kind of argument made sense. Today, it'due south much less valid. Engineering advances year-on-twelvemonth, just the rate and footstep of those advances tin vary significantly.

Polaris2

This slide is by AMD, but the advantages of FinFET over 28nm planar transistors use to both companies.

The 14/16nm node is a major stepping stone for GPU performance because it's the first full-node shrink that's been available to the GPU manufacture in more than four years. If y'all care about low power consumption and small class factors, upcoming chips should be dramatically more power efficient. If y'all care about high-end operation, yous may take to expect another ix months, but the corporeality of GPU you'll be able to buy for the same amount of money should be xxx-50% higher than what y'all'll get today.

There's besides the question of VR engineering science. We don't know yet how VR will evolve or how seriously it will touch on the futurity of gaming; estimates I've seen range from full transformation to a niche marketplace for a handful of well-heeled enthusiasts. Regardless, if you lot plan on jumping on the VR bandwagon, information technology behooves you to wait and see what kind of operation next-generation video cards tin can offering.

Remember this: VR applied science demands both high frame rates and extremely smooth frame delivery, and this has knock-on effects on which GPUs can reliably deliver that experience. A GPU that drives 50 frames per second where 30 is a minimum requirement is pushing 1.67x more frames per second than the user demands as a minimum standard. A GPU that delivers 110 frames per second where ninety is a minimum requirement is only 1.22x higher up the target frame rate. Information technology doesn't take much in the way of boosted heart candy earlier our second GPU is bottoming out at ninety FPS again.

The concluding reason to consider delaying an upgrade is whether you plan to upgrade to a 4K monitor at whatever signal in the adjacent few years. 4K pushes roughly 4x more pixels than 1080p monitors and modern graphics cards are often 33-50% slower when gaming at that resolution. Waiting a few more months to buy at the beginning of the new wheel could hateful fifty% more than performance for the aforementioned price and gives yous a ameliorate run a risk of ownership a carte that can handle 4K in a wider variety of titles.

If your GPU suddenly dies tomorrow or you can't stand running an sometime HD 5000 or GTX 400-series card another minute, you tin upgrade to a newer AMD or Nvidia model available today and all the same run into an enormous functioning uplift — simply customers who can wait for the side by side-generation refreshes to arrive volition be getting much more than blindside for their cadet. We don't know what the verbal specs volition be for any specific AMD or Nvidia next-gen GPU, but what we're seeing and hearing about the 16/14nm node is extremely encouraging. If you tin wait, you nearly certainly won't regret it — especially if you want a clearer pic on which company, AMD or Nvidia, performs amend in DirectX 12.