Feeds

Nvidia snaps out snappier Tesla GPU coprocessors

All fired up on all 512 cores

Internet Security Threat Report 2014

Do the math

When you add it up, the faster clocks and memory and the higher core count add up to somewhere between a 20 and 30 per cent improvement in performance, thus:

Nvidia Tesla M2090 performance

Tesla M2090: Your performance improvement may vary

By the raw numbers, the M2090 is rated at 665 gigaflops at double-precision and 1.33 teraflops at single-precision, or 29.1 per cent more than the M2070 it replaces. The M2090 delivers 178GB/sec of memory bandwidth, up from 148GB/sec with the M2070.

And here's the kicker: the M2090s go faster but are still within the same 225 watt peak power draw of the earlier and slower M2050 and M2070 devices. Of course, your actual power draw will depend on the workload and how it stresses the GPU coprocessor.

But its not just speed and power requirements that are pushing the coprocessor envelope: the big innovation that is helping with the adoption of GPU accelerators for supercomputing and other analytics jobs is not the GPU, but the server that wraps around them.

"We're beginning to see server makers respond to what customers want," Gupta tells El Reg. "Once customers start using GPUs, they want more GPUs in a box and fewer CPUs."

Gupta calls out HP's new ProLiant SL390s G7, which can cram eight GPUs into a two-socket tray server in a half-width 4U tray, which was soft-launched back in April. The Tesla M2050 and M2070 GPU coprocessors were already certified in this machine, as is the new M2090. That full configuration gives customers a ratio of four GPUs per socket – but, oddly enough, this is not actually good enough. For most HPC applications, one GPU per CPU core is what the applications really need, says Gupta.

On a two-socket server, if you backstep from a six-core Xeon 5600 to a four-core model (such as the Xeon X5667 launched in February), you can get things into balance. At least until there is an eight-core Xeon chip.

Perhaps the better answer, you're thinking, might be to do what Nvidia has already done for workstation graphics back at the end of March, and double-up the GPUs on a double-wide card.

The GTX 590 graphics card has two 512-core Fermi chips running at 1.22GHz with 1.5GB of GDDR5 memory per chip running at 1.77GHz. Nvidia does not provide floating-point performance ratings on the GTX 590, but it should be somewhere around 1.24 teraflops double-precision. With six of these GTX 590 cards in a 3U chassis, an HP ProLiant SL390s would be able to use six-core Xeon 5600 chips and have one GPU per CPU core.

The only trouble is, however, that the GTX 590s are aimed at workstations, not servers, and have fans on them that might mess up the airflow in the server chassis. Moreover, the GTX 590 cards pull down 365 watts of power, and are therefore a bit more power dense than the M2090, which has half as many GPUs but four times as much GDDR5 memory per Fermi chip. If your HPC app needs more GPU and less GPU memory, this is worth thinking about, if your server can handle the cooling job on multiple GTX 590s – and you can actually get your hands on them.

The GTX 590 will win, hands down, on price, at $699. The M2090 is probably going to cost somewhere around $4,100 to $4,700, if it is priced like the M2070 it replaces. Nvidia does not provide official pricing on the fanless M20 series of GPU coprocessors, which is silly.

Last week, Nvidia CEO Jen-Hsun Huang lamented that the Professional Solutions business, where workstation graphics and Tesla co-processors live inside Nvidia, was not doing as well as expected, and part of the problem would seem to be that Nvidia is trying to charge too much money for the fanless GPUs. The price disparity between the GTX 590 and the M2090 is too large, plain and simple. ®

Internet Security Threat Report 2014

More from The Register

next story
Docker's app containers are coming to Windows Server, says Microsoft
MS chases app deployment speeds already enjoyed by Linux devs
IBM storage revenues sink: 'We are disappointed,' says CEO
Time to put the storage biz up for sale?
'Hmm, why CAN'T I run a water pipe through that rack of media servers?'
Leaving Las Vegas for Armenia kludging and Dubai dune bashing
'Urika': Cray unveils new 1,500-core big data crunching monster
6TB of DRAM, 38TB of SSD flash and 120TB of disk storage
Facebook slurps 'paste sites' for STOLEN passwords, sprinkles on hash and salt
Zuck's ad empire DOESN'T see details in plain text. Phew!
SDI wars: WTF is software defined infrastructure?
This time we play for ALL the marbles
Windows 10: Forget Cloudobile, put Security and Privacy First
But - dammit - It would be insane to say 'don't collect, because NSA'
prev story

Whitepapers

Forging a new future with identity relationship management
Learn about ForgeRock's next generation IRM platform and how it is designed to empower CEOS's and enterprises to engage with consumers.
Cloud and hybrid-cloud data protection for VMware
Learn how quick and easy it is to configure backups and perform restores for VMware environments.
Three 1TB solid state scorchers up for grabs
Big SSDs can be expensive but think big and think free because you could be the lucky winner of one of three 1TB Samsung SSD 840 EVO drives that we’re giving away worth over £300 apiece.
Reg Reader Research: SaaS based Email and Office Productivity Tools
Read this Reg reader report which provides advice and guidance for SMBs towards the use of SaaS based email and Office productivity tools.
Security for virtualized datacentres
Legacy security solutions are inefficient due to the architectural differences between physical and virtual environments.