- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
Hello,
I would like to port a GCN-optimized 1.2 kernels to run on latest Intel GPUs.
Are there any general guidelines I should be following?
For example, on GCN, if work group size is smaller than 64 (wave front size), then it is possible to dispense with
memory barriers, since work items will never be executed more than one Compute Unit. Does this apply to Intel GPU?
Any other things to keep in mind?
Also, in terms of install base, what version of HD Graphics GPU is the most common in the field?
And which CPUs have these GPUs?
Thanks so much,
Aaron
Link Copied
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
Aaron,
It would be good to review the following documents before you get started:
https://software.intel.com/sites/default/files/managed/63/2c/SF14_GVCS002_100f.pdf
The most common graphics configurations are GT2s with 16 EUs (execution units == compute units) on Ivy Bridge (3rd generation Intel(R) Core Processors), 20 EUs on Haswell (4th generation Intel(R) Core Processor), 24 EUs on Broadwell (5th gen Core), 24 EUs on Skylake (6th Gen Core). As far as I know, the most common chips out there are still 3rd gen processors with Intel® HD Graphics 4000 (16 EUs): see http://ark.intel.com/search?q=HD+Graphics+4000 - see also http://store.steampowered.com/hwsurvey and http://store.steampowered.com/hwsurvey
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
Great, thanks a lot!

- Subscribe to RSS Feed
- Mark Topic as New
- Mark Topic as Read
- Float this Topic for Current User
- Bookmark
- Subscribe
- Printer Friendly Page