-----Original Message----- From: Boost [mailto:boost-bounces@lists.boost.org] On Behalf Of Jason Newton Sent: Thursday, January 01, 2015 4:46 To: boost@lists.boost.org Subject: Re: [boost] [compute] Review
I'd be very interested in seeing a valid and direct comparison of a few common algorithms implemented in the same fashion in OpenCL and CUDA in addition library-to-library performance charts [here: https://kylelutz.github.io/compute/boost_compute/performance.html] Boost.Compute has already - this might take away some of the fear from the research community of OpenCL vs CUDA esp given the strength of the Boost name.
The biggest issue I see with those performance comparisons is that they're not very realistic. While a direct comparisons of individual operation performance is certainly interesting, a more relevant metric would be to implement and compare solutions to real-world problems. This will expose various copying and synchronization overheads, which are doubtlessly of concern to many current GPGPU practitioners and potential Boost.Compute users.
A few tidbits relating to the work other than that though: -One of the things I don't like is the scan operation recursively calls itself and reallocates memory a bunch of times - Bolt seems to have taken a better approach here which reduces the global memory usage too. I don't like additional memory allocations or things that could incur unnecessary/unexpected latencies if the problem changes size.
At the very least, documentation of the memory requirements might help users partition their data appropriately to avoid device memory exhaustion. Matt ________________________________ This e-mail contains privileged and confidential information intended for the use of the addressees named above. If you are not the intended recipient of this e-mail, you are hereby notified that you must not disseminate, copy or take any action in respect of any information contained in it. If you have received this e-mail in error, please notify the sender immediately by e-mail and immediately destroy this e-mail and its attachments.