I've found a few bugs in the Intel Intrinsics Guide 2.7 (I'm using Linux version):
1. When the window is maximized, the search field is stretched vertically while still being a one-line edit box. It sould probably be sized accordingly.
2. __m256 _mm256_undefined_si256 () should return __m256i.
3. In some instructions description, like _mm_adds_epi8, the operation is described in terms of SignedSaturate while, e.g. _mm256_adds_epi16 is described with SaturateToSignedWord. This applies to other operations with unsigned saturation as well. Also, the vector elements are described differently. More consistent description would be nice.
4. _mm_alignr_epi8 has two descriptions.
5. I'm not sure _mm_ceil_pd signature and description is correct. It says the intrinsic returns a vector of single-precision floats. Shouldn't it be double-precision?
I didn't read all instructions so there may be more issues. I'll post if I find anything else.
PS: This is not a bug per se but some instructions are missing the Latency & Throughput information. This mostly relates to newer instructions but still this info is useful and I hope it will be added.
Still doesn't work today, it seems that no one is in charge.
Osiv, Oleksiy wrote:
Hey, the guide is not working at all today. I checked Chrome & Edge. Development console contains the following error:
Refused to execute script from 'https://software.intel.com/sites/landingpage/IntrinsicsGuide/files/perf....' because its MIME type ('application/json') is not executable, and strict MIME type checking is enabled.
Hello.... Intel developers...!
Hello, Intel developers,
What does multiplication of scaled index by constant 8 mean in description of _mm512_i64gather_ps (and similar gather/scatter functions)? As far as i can observe, actual behavior of such functions does not include the additional constant scaling.
In various gather operation descriptions for AVX2 and AVX-512, the scale is multiplied by 8 (`ZeroExtend64(scale) * 8`). This is incorrect, as the scale is the actual multiplier, without multiplying it by 8. This follows from SDM description of VPGATHERDD/DQ/QD/QQ.
The Intrinsics Guide seems to be down. The page says "Error Loading Data" and the search doesn't work.
Also, the link in the tooltip displayed when you hover over the question mark "?" is outdated: http://software.intel.com/en-us/forums/topic/363747. This link used to point to this thread in the old forum and now leads to some general hub of Intel forums.
A better yet solution would be if Intel created a proper bug tracker to report issues in the Intrinsics Guide. Like, if you hosted the project on GitHub or something.
To add to the "Error Loading Data" error, web browser console has these errors:
Cross-Origin Request Blocked: The Same Origin Policy disallows reading the remote resource at https://corpredirect.intel.com/Redirector/404Redirector.aspx?https://software.intel.com/en-us/intel_.... (Reason: CORS header ‘Access-Control-Allow-Origin’ missing).
Cross-Origin Request Blocked: The Same Origin Policy disallows reading the remote resource at https://corpredirect.intel.com/Redirector/404Redirector.aspx?https://software.intel.com/en-us/intel_.... (Reason: CORS request did not succeed).