<html><head><meta http-equiv="Content-Type" content="text/html; charset=utf-8"></head><body style="word-wrap: break-word; -webkit-nbsp-mode: space; line-break: after-white-space;" class="">Hi,<div class=""><br class=""><div><br class=""><blockquote type="cite" class=""><div class="">On Aug 12, 2020, at 12:25, Sjoerd Meijer <<a href="mailto:Sjoerd.Meijer@arm.com" class="">Sjoerd.Meijer@arm.com</a>> wrote:</div><br class="Apple-interchange-newline"><div class=""><div style="font-style: normal; font-variant-caps: normal; font-weight: normal; letter-spacing: normal; text-align: start; text-indent: 0px; text-transform: none; white-space: normal; word-spacing: 0px; -webkit-text-stroke-width: 0px; text-decoration: none; font-family: Calibri, Arial, Helvetica, sans-serif; font-size: 12pt;" class="">Hi Florian (+cfe-dev for visibilty),</div><div style="font-style: normal; font-variant-caps: normal; font-weight: normal; letter-spacing: normal; text-align: start; text-indent: 0px; text-transform: none; white-space: normal; word-spacing: 0px; -webkit-text-stroke-width: 0px; text-decoration: none; font-family: Calibri, Arial, Helvetica, sans-serif; font-size: 12pt;" class=""><br class=""></div><div style="font-style: normal; font-variant-caps: normal; font-weight: normal; letter-spacing: normal; text-align: start; text-indent: 0px; text-transform: none; white-space: normal; word-spacing: 0px; -webkit-text-stroke-width: 0px; text-decoration: none; font-family: Calibri, Arial, Helvetica, sans-serif; font-size: 12pt;" class="">I was playing a little bit with Clang matrix language extension and and wanted to check with you to see if I am not missing something about the matrix type conversion. The draft spec says:</div><div style="font-style: normal; font-variant-caps: normal; font-weight: normal; letter-spacing: normal; text-align: start; text-indent: 0px; text-transform: none; white-space: normal; word-spacing: 0px; -webkit-text-stroke-width: 0px; text-decoration: none; font-family: Calibri, Arial, Helvetica, sans-serif; font-size: 12pt;" class=""><br class=""></div><div style="font-style: normal; font-variant-caps: normal; font-weight: normal; letter-spacing: normal; text-align: start; text-indent: 0px; text-transform: none; white-space: normal; word-spacing: 0px; -webkit-text-stroke-width: 0px; text-decoration: none; font-family: Calibri, Arial, Helvetica, sans-serif; font-size: 12pt;" class="">"A value of matrix type can be converted to another matrix type if the number of rows and columns are the same and the value’s elements can be converted to the element type of the result type. "</div><div style="font-style: normal; font-variant-caps: normal; font-weight: normal; letter-spacing: normal; text-align: start; text-indent: 0px; text-transform: none; white-space: normal; word-spacing: 0px; -webkit-text-stroke-width: 0px; text-decoration: none; font-family: Calibri, Arial, Helvetica, sans-serif; font-size: 12pt;" class=""><br class=""></div><div style="font-style: normal; font-variant-caps: normal; font-weight: normal; letter-spacing: normal; text-align: start; text-indent: 0px; text-transform: none; white-space: normal; word-spacing: 0px; -webkit-text-stroke-width: 0px; text-decoration: none; font-family: Calibri, Arial, Helvetica, sans-serif; font-size: 12pt;" class="">I have tried a different variants of this:</div><div style="font-style: normal; font-variant-caps: normal; font-weight: normal; letter-spacing: normal; text-align: start; text-indent: 0px; text-transform: none; white-space: normal; word-spacing: 0px; -webkit-text-stroke-width: 0px; text-decoration: none; font-family: Calibri, Arial, Helvetica, sans-serif; font-size: 12pt;" class=""><br class=""></div><div style="font-style: normal; font-variant-caps: normal; font-weight: normal; letter-spacing: normal; text-align: start; text-indent: 0px; text-transform: none; white-space: normal; word-spacing: 0px; -webkit-text-stroke-width: 0px; text-decoration: none; font-family: Calibri, Arial, Helvetica, sans-serif; font-size: 12pt;" class=""> typedef char m2x8_t __attribute__((matrix_type(2, 8)));<div class=""> typedef char m8x2_t __attribute__((matrix_type(8, 2)));</div><div class=""> typedef char m2x2_char_t __attribute__((matrix_type(2, 2)));</div><div class=""> typedef int m2x2_int_t __attribute__((matrix_type(2, 2)));</div><div class=""><br class=""></div><div class=""> m2x2_int_t f(m2x8_t a, m8x2_t b) {</div><div class=""> return static_cast<m2x2_int_t>(a *b);</div><div class=""> }</div><div class=""><br class=""></div><div class="">but am getting errors that this conversion is not allowed. Unless I am doing something very silly here, I am guessing that this because the matrix extension is work in progress?</div><div class=""><br class=""></div></div></div></blockquote><div><br class=""></div>This should work according to the spec, but the conversion has not been implemented yet I think. I’ve created <a href="https://bugs.llvm.org/show_bug.cgi?id=47141" class="">https://bugs.llvm.org/show_bug.cgi?id=47141</a> and linked it to <a href="https://bugs.llvm.org/show_bug.cgi?id=46163" class="">https://bugs.llvm.org/show_bug.cgi?id=46163</a> which should act as an umbrella issue to track the missing pieces.</div><div><br class=""></div><div><br class=""></div><div><blockquote type="cite" class=""><div class=""><div style="font-style: normal; font-variant-caps: normal; font-weight: normal; letter-spacing: normal; text-align: start; text-indent: 0px; text-transform: none; white-space: normal; word-spacing: 0px; -webkit-text-stroke-width: 0px; text-decoration: none; font-family: Calibri, Arial, Helvetica, sans-serif; font-size: 12pt;" class=""><div class="">The draft spec also says that implicit conversions don't apply, but that would perhaps be convenient? But I haven't given this any thoughts yet though if that could be problematic.</div></div></div></blockquote><div><br class=""></div>I think currently we match the behavior for vector types and only convert scalar operands for binary operators implicitly to matrixes. If there’s a strong need for implicit conversions, this is certainly something that can be revisited.</div><div><br class=""></div><div><blockquote type="cite" class=""><div class=""><div style="font-style: normal; font-variant-caps: normal; font-weight: normal; letter-spacing: normal; text-align: start; text-indent: 0px; text-transform: none; white-space: normal; word-spacing: 0px; -webkit-text-stroke-width: 0px; text-decoration: none; font-family: Calibri, Arial, Helvetica, sans-serif; font-size: 12pt;" class=""><div class=""><br class=""></div><div class="">Moving on a bit to lowering this to te matrix multiply intrinsics. I think it would be convenient if the matrix multiply can accumulate in a wider type (because that's what some instructions do). While there are probably different approaches possible, the llvm intrinsic has the vector type for the return value and its arguments:</div><div class=""><pre style="overflow: auto hidden; font-family: Consolas, "Deja Vu Sans Mono", "Bitstream Vera Sans Mono", monospace; font-size: 0.95em; line-height: 15.96px; padding: 0.5em; border: 1px solid rgb(204, 204, 204); background-color: rgb(248, 248, 248); text-align: left;" class=""><span class="n">vectorty</span> <span class="nd" style="color: rgb(85, 85, 85); font-weight: bold;">@llvm</span><span class="o" style="color: rgb(102, 102, 102);">.</span><span class="n">matrix</span><span class="o" style="color: rgb(102, 102, 102);">.</span><span class="n">multiply</span><span class="o" style="color: rgb(102, 102, 102);">.*</span><span class="p">(</span><span class="n">vectorty</span> <span class="o" style="color: rgb(102, 102, 102);">%</span><span class="n">A</span><span class="p">,</span> <span class="n">vectorty</span> <span class="o" style="color: rgb(102, 102, 102);">%</span><span class="n">B</span><span class="p">, ...)</span></pre> So perhaps we can relax this?</div></div></div></blockquote><br class=""></div><div><br class=""></div><div>Yes we can certainly extend this, to allow use cases to map to hardware instructions that implement an extension step, like AAch64’s udot. IIRC it extends the sums, which I think would make the most sense to use, as otherwise it should be sufficient to extend the operands/result.</div><div><br class=""></div><div>I think the more interesting question here would be how this fits into the C/C++ spec. I guess it would be possible to specify it so a multiply that gets extended lowers to the widening intrinsic, but this would seem quite surprising/awkward. In my opinion, a separate builtin would be a cleaner solution.</div><div><br class=""></div><div>Cheers,</div><div>Florian</div><br class=""></div></body></html>