CUDA woulda shoulda...did
In what can only be described as a Far Cry from NVISION 2008, NVIDIA's GPU Technology Conference (GTC) 2009 was a strictly grown up affair.
Gone were the greasy-haired, red-bull-fuelled, cold-pizza-filled kids - propping their eyelids open with matchsticks to finish 72 hours of consecutive LAN party gaming. In were the suited and shiny-booted bespectacled academics and business men with briefcases and white boards filled with mathematical equations. The tech press mumbled in disbelief, some indignant, others admiring; NVIDIA seemed to have grown up.
Before the show, many had posited that NVIDIA was a company in Crysis, with Intel and AMD forming what can only be described as an Assassin's Creed, attempting to cut off the firm's future with fusion CPU/GPU offerings and perhaps, a few years down the line, Larrabee.
But realising that if it didn't move fast, the firm would perhaps be relegated to something of a Dead Space in the PC industry, NVIDIA seems to have pulled itself up by its bat breaches and created a developer conference of enviable buzz and excitement, even if it was a little weak on newsworthy hardware announcements. It even managed to make Intel's Developer Forum feel slightly anaemic in comparison.
Thousands of developers vied to cram themselves into sessions which had been overbooked for weeks, all straining to see highly-complex-looking presentations on spare linear algebra, quantum chemistry on GPU, computational finance, fluid dynamics and binary black hole evolution. Bioshocked? We were.
Because, as Nvidia repeatedly pointed out, whilst AMD stands in a corner hurling abuse at its ‘proprietary CUDA' and telling the world how Open CL - when it eventually appears in the real world - will leave it for dead, NVIDIA itself actually seems to be doing something useful. And developers, academics, scientists and financial number crunchers are lapping it up.
"Who cares about transcoding and encoding your crappy mobile phone videos to YouTube in 34 seconds instead of 50; we care about finding the cure for cancer" was NVIDIA's message this year, and it's one that must be leaving AMD feeling rather childish and immature.
Of course, we're well aware that NVIDIA is using a rather clever marketing ploy, rather than heroically answering science's Call of Duty, but Kudos (Cudas?) to the firm for trying, anyway.