NVIDIA Update: CUDA Week in Review (CUDA Pro Tip; Spotlight on Speech Recognition)

CUDA PRO TIP The new Maxwell GPU architecture is the most power efficient GPU yet. Maxwell improves instruction scheduling and latency, and adds more dedicated shared memory and shared memory atomic instructions. Learn all about it on the Parallel Forall Blog.

GPU TECH CONFERENCE

Keynotes Revealed Beyond the deep-dive technical sessions, interactive tutorials and hands-on labs, we’ve assembled an all-star roster of keynotes to offer insight and ignite your imagination. Here’s the lineup:   Jen-Hsun Huang, NVIDIA CEO and Co-Founder (opening keynote on March 25)   Dirk Van Gelder and Danny Nahmias of Pixar (March 26)   Adam Gazzaley, visionary neuroscientist, UC San Francisco (March 27) Looking forward to seeing you soon in San Jose, Calif. (20% discount code: GM20CD)

CUDA SPOTLIGHT
CUDA-Accelerated Speech Recognition Our Spotlight is on Ian Lane of Carnegie Mellon University. Dr. Lane leads the speech and language processing group at CMU Silicon Valley. His team is developing methods to accelerate speech and language technologies using GPUs. Read the Spotlight.
CUDA NEWS
CUDA 6 RC Now Available to All CUDA 6 Release Candidate (RC) is now available to all, featuring Unified Memory; Drop-in Libraries; and Multi-GPU Scaling in cuBLAS and cuFFT Libraries. Downloadtoday and check out these upcoming webinars: March 11: CUDA 6: Performance Review March 13: CUDA 6: Drop-in Performance Optimized Libraries March 18: CUDA 6: Unified Memory
GPU MEETUPS
March 25: Silicon Valley (Co-located with GTC14. All GPU Meetup members are welcome) Special guest speaker: Dr. JoAnn Kuchera-Morin, University of California, Santa Barbara