<html><head><meta http-equiv="Content-Type" content="text/html charset=windows-1252"></head><body style="word-wrap: break-word; -webkit-nbsp-mode: space; -webkit-line-break: after-white-space;" class=""><br class=""><div><blockquote type="cite" class=""><div class="">On Aug 25, 2014, at 4:43 PM, Quentin Colombet <<a href="mailto:qcolombet@apple.com" class="">qcolombet@apple.com</a>> wrote:</div><br class="Apple-interchange-newline"><div class=""><span style="font-family: Helvetica; font-size: 12px; font-style: normal; font-variant: normal; font-weight: normal; letter-spacing: normal; line-height: normal; orphans: auto; text-align: start; text-indent: 0px; text-transform: none; white-space: normal; widows: auto; word-spacing: 0px; -webkit-text-stroke-width: 0px; float: none; display: inline !important;" class="">Thanks Bob for the context.</span><br style="font-family: Helvetica; font-size: 12px; font-style: normal; font-variant: normal; font-weight: normal; letter-spacing: normal; line-height: normal; orphans: auto; text-align: start; text-indent: 0px; text-transform: none; white-space: normal; widows: auto; word-spacing: 0px; -webkit-text-stroke-width: 0px;" class=""><br style="font-family: Helvetica; font-size: 12px; font-style: normal; font-variant: normal; font-weight: normal; letter-spacing: normal; line-height: normal; orphans: auto; text-align: start; text-indent: 0px; text-transform: none; white-space: normal; widows: auto; word-spacing: 0px; -webkit-text-stroke-width: 0px;" class=""><span style="font-family: Helvetica; font-size: 12px; font-style: normal; font-variant: normal; font-weight: normal; letter-spacing: normal; line-height: normal; orphans: auto; text-align: start; text-indent: 0px; text-transform: none; white-space: normal; widows: auto; word-spacing: 0px; -webkit-text-stroke-width: 0px; float: none; display: inline !important;" class="">On Aug 25, 2014, at 4:05 PM, Bob Wilson <</span><a href="mailto:bob.wilson@apple.com" style="font-family: Helvetica; font-size: 12px; font-style: normal; font-variant: normal; font-weight: normal; letter-spacing: normal; line-height: normal; orphans: auto; text-align: start; text-indent: 0px; text-transform: none; white-space: normal; widows: auto; word-spacing: 0px; -webkit-text-stroke-width: 0px;" class="">bob.wilson@apple.com</a><span style="font-family: Helvetica; font-size: 12px; font-style: normal; font-variant: normal; font-weight: normal; letter-spacing: normal; line-height: normal; orphans: auto; text-align: start; text-indent: 0px; text-transform: none; white-space: normal; widows: auto; word-spacing: 0px; -webkit-text-stroke-width: 0px; float: none; display: inline !important;" class="">> wrote:</span><br style="font-family: Helvetica; font-size: 12px; font-style: normal; font-variant: normal; font-weight: normal; letter-spacing: normal; line-height: normal; orphans: auto; text-align: start; text-indent: 0px; text-transform: none; white-space: normal; widows: auto; word-spacing: 0px; -webkit-text-stroke-width: 0px;" class=""><br style="font-family: Helvetica; font-size: 12px; font-style: normal; font-variant: normal; font-weight: normal; letter-spacing: normal; line-height: normal; orphans: auto; text-align: start; text-indent: 0px; text-transform: none; white-space: normal; widows: auto; word-spacing: 0px; -webkit-text-stroke-width: 0px;" class=""><blockquote type="cite" style="font-family: Helvetica; font-size: 12px; font-style: normal; font-variant: normal; font-weight: normal; letter-spacing: normal; line-height: normal; orphans: auto; text-align: start; text-indent: 0px; text-transform: none; white-space: normal; widows: auto; word-spacing: 0px; -webkit-text-stroke-width: 0px;" class="">+ Jim Grosbach<br class=""><br class=""><blockquote type="cite" class="">On Aug 25, 2014, at 10:45 AM, Quentin Colombet <<a href="mailto:qcolombet@apple.com" class="">qcolombet@apple.com</a>> wrote:<br class=""><br class="">Thanks Renato.<br class=""><br class="">I’ll wait for Bob’s inputs but I think, like you, removing the optimization level is the right thing to do.<br class=""><br class="">-Quentin<br class=""></blockquote><br class="">This is kind of an unusual test that doesn’t fit the usual mold. I’m hesitant to defend it, except to say that so far, it is better than the alternatives.<br class=""><br class="">Jim and I have debated this at some length, and Jim agrees with both of you that the “right” thing to do is to have a front-end test that checks the IR output of the front-end, with separate tests in the backend to check that the IR is translated to the expected assembly.<br class=""><br class="">We don’t have that.<br class=""><br class="">We have backend tests for some Neon code-gen and perhaps a few front-end tests for intrinsics as well, but there are a huge number of Neon intrinsics and it is important to have exhaustive tests for all of them. That’s what this test is for. It is machine-generated from clang’s arm_neon.td file. Obviously, if someone introduces a bug in the .td file, we would not catch it with this test. To avoid that, we have checked-in this pre-generated test, which we have manually checked against ARM’s documentation of the Neon intrinsics. I had requested an automatic test to regenerate it and check for differences, but I’m not sure that has gotten implemented yet.<br class=""><br class="">So, anyway, that is the context. The reason it is compiled with optimization is that the auto-generated checks need to look for certain opcodes and without optimization, you won’t get many of them.<br class=""></blockquote><br style="font-family: Helvetica; font-size: 12px; font-style: normal; font-variant: normal; font-weight: normal; letter-spacing: normal; line-height: normal; orphans: auto; text-align: start; text-indent: 0px; text-transform: none; white-space: normal; widows: auto; word-spacing: 0px; -webkit-text-stroke-width: 0px;" class=""><span style="font-family: Helvetica; font-size: 12px; font-style: normal; font-variant: normal; font-weight: normal; letter-spacing: normal; line-height: normal; orphans: auto; text-align: start; text-indent: 0px; text-transform: none; white-space: normal; widows: auto; word-spacing: 0px; -webkit-text-stroke-width: 0px; float: none; display: inline !important;" class="">What do you recommend for the test cases where we are able to optimize the sequence of instructions? (See my examples form a previous mail)</span><br style="font-family: Helvetica; font-size: 12px; font-style: normal; font-variant: normal; font-weight: normal; letter-spacing: normal; line-height: normal; orphans: auto; text-align: start; text-indent: 0px; text-transform: none; white-space: normal; widows: auto; word-spacing: 0px; -webkit-text-stroke-width: 0px;" class=""></div></blockquote><div><br class=""></div>Well, as you can see, we just omit the checks. There weren’t a lot of other options when we were auto generating the tests from very limited info in the arm_neon.td file. I should also point out that when James Molloy rewrote the TableGen Neon intrinsic emitter in r211101, he removed all the testcase generation. I don’t know the status of putting it back. James, is that something that even makes sense to do now?</div><div><br class=""></div><div><blockquote type="cite" class=""><div class=""><span style="font-family: Helvetica; font-size: 12px; font-style: normal; font-variant: normal; font-weight: normal; letter-spacing: normal; line-height: normal; orphans: auto; text-align: start; text-indent: 0px; text-transform: none; white-space: normal; widows: auto; word-spacing: 0px; -webkit-text-stroke-width: 0px; float: none; display: inline !important;" class="">Since we have to run with the optimizations enabled, is it worth reworking the test cases to use the values produced by the copy-related intrinsics?</span><br style="font-family: Helvetica; font-size: 12px; font-style: normal; font-variant: normal; font-weight: normal; letter-spacing: normal; line-height: normal; orphans: auto; text-align: start; text-indent: 0px; text-transform: none; white-space: normal; widows: auto; word-spacing: 0px; -webkit-text-stroke-width: 0px;" class=""><span style="font-family: Helvetica; font-size: 12px; font-style: normal; font-variant: normal; font-weight: normal; letter-spacing: normal; line-height: normal; orphans: auto; text-align: start; text-indent: 0px; text-transform: none; white-space: normal; widows: auto; word-spacing: 0px; -webkit-text-stroke-width: 0px; float: none; display: inline !important;" class="">That way the moves wouldn’t be coalesced.</span><br style="font-family: Helvetica; font-size: 12px; font-style: normal; font-variant: normal; font-weight: normal; letter-spacing: normal; line-height: normal; orphans: auto; text-align: start; text-indent: 0px; text-transform: none; white-space: normal; widows: auto; word-spacing: 0px; -webkit-text-stroke-width: 0px;" class=""></div></blockquote><div><br class=""></div>I suppose. Since we no longer have the automatic test generation, maybe we should just use the existing test as a starting point and go ahead and make incremental improvements to it. I’m not opposed to that, anyway.</div><div><br class=""><blockquote type="cite" class=""><div class=""><br style="font-family: Helvetica; font-size: 12px; font-style: normal; font-variant: normal; font-weight: normal; letter-spacing: normal; line-height: normal; orphans: auto; text-align: start; text-indent: 0px; text-transform: none; white-space: normal; widows: auto; word-spacing: 0px; -webkit-text-stroke-width: 0px;" class=""><span style="font-family: Helvetica; font-size: 12px; font-style: normal; font-variant: normal; font-weight: normal; letter-spacing: normal; line-height: normal; orphans: auto; text-align: start; text-indent: 0px; text-transform: none; white-space: normal; widows: auto; word-spacing: 0px; -webkit-text-stroke-width: 0px; float: none; display: inline !important;" class="">Thanks,</span><br style="font-family: Helvetica; font-size: 12px; font-style: normal; font-variant: normal; font-weight: normal; letter-spacing: normal; line-height: normal; orphans: auto; text-align: start; text-indent: 0px; text-transform: none; white-space: normal; widows: auto; word-spacing: 0px; -webkit-text-stroke-width: 0px;" class=""><span style="font-family: Helvetica; font-size: 12px; font-style: normal; font-variant: normal; font-weight: normal; letter-spacing: normal; line-height: normal; orphans: auto; text-align: start; text-indent: 0px; text-transform: none; white-space: normal; widows: auto; word-spacing: 0px; -webkit-text-stroke-width: 0px; float: none; display: inline !important;" class="">-Quentin<span class="Apple-converted-space"> </span></span><br style="font-family: Helvetica; font-size: 12px; font-style: normal; font-variant: normal; font-weight: normal; letter-spacing: normal; line-height: normal; orphans: auto; text-align: start; text-indent: 0px; text-transform: none; white-space: normal; widows: auto; word-spacing: 0px; -webkit-text-stroke-width: 0px;" class=""><br style="font-family: Helvetica; font-size: 12px; font-style: normal; font-variant: normal; font-weight: normal; letter-spacing: normal; line-height: normal; orphans: auto; text-align: start; text-indent: 0px; text-transform: none; white-space: normal; widows: auto; word-spacing: 0px; -webkit-text-stroke-width: 0px;" class=""><blockquote type="cite" style="font-family: Helvetica; font-size: 12px; font-style: normal; font-variant: normal; font-weight: normal; letter-spacing: normal; line-height: normal; orphans: auto; text-align: start; text-indent: 0px; text-transform: none; white-space: normal; widows: auto; word-spacing: 0px; -webkit-text-stroke-width: 0px;" class=""><br class=""><blockquote type="cite" class="">On Aug 25, 2014, at 10:31 AM, Renato Golin <<a href="mailto:renato.golin@linaro.org" class="">renato.golin@linaro.org</a>> wrote:<br class=""><br class=""><blockquote type="cite" class="">On 25 August 2014 17:57, Quentin Colombet <<a href="mailto:qcolombet@apple.com" class="">qcolombet@apple.com</a>> wrote:<br class=""><blockquote type="cite" class="">The problem is that the intrinsics at stake are just fancy moves, that can<br class="">be coalesced. I do not think there is a way to prevent the optimization to<br class="">happen other than disabling the optimization.<br class=""></blockquote><br class="">Well, those intrinsics are useful for a few things, normally a<br class="">sequence of neon calls to which the move is not clear and needs to be<br class="">explicit. Although this case was obvious and "easy" to generate, it's<br class="">normally the first optimizations you end up doing, so not stable<br class="">enough.<br class=""><br class=""><br class=""><blockquote type="cite" class="">I can add the flag to do that, but I guess that wouldn’t be the right fix,<br class="">since we could have another backend that LLVM.<br class="">If we do want to check the lowering of intrinsics, shouldn’t we drop Os from<br class="">the run command?<br class=""></blockquote><br class="">Yes! We want to test if the front-end is generating the correct<br class="">intrinsic, so any level of optimization will hide whatever we were<br class="">trying to test in the first place. I risk to say that whomever removed<br class="">the other tests in this file also did wrong, and we need to fix this.<br class=""><br class="">I'm copying Bob since he was the one writing the NEON intrinsics in<br class="">the first place, maybe he can see things I fail to. But AFAICS,<br class="">disabling *any* optimization and re-checking for the correct<br class="">instructions is the way to go.<br class=""><br class="">cheers,<br class="">--renato</blockquote></blockquote></blockquote></div></blockquote></div><br class=""></body></html>