similar to: Side-channel resistant values

On 9/12/19 5:06 AM, David Zarzycki via llvm-dev wrote: I think adding a builtin to force CMOV or similar instructions on other architectures is long overdue. It’s generally useful, even if one isn’t mitigating speculative execution. I believe that you can currently get this effect using __builtin_unpredictable in Clang. __builtin_unpredictable wasn't added for this purpose, and it's a

Hi Chandler, The data-invariant feature sounds great but what about the general case? When performance tuning code, people sometimes need the ability to reliably generate CMOV, and right now the best advice is either “use inline assembly” or “keep refactoring until CMOV is emited” (and hope that future compilers continue to generate CMOV). Given that a patch already exists to reliably generate

Dear R-users & developers, i'd like to know your opinion on the following suggestion. In short: Wouldn't it be possible to combine the properties of local regression (loess) and resistant regression (lqs) in order to cope with fitting of partially continuous data? I think the lqs-fitting could detect and eliminate the 2 outliers but perserve the discontinuity at t1 in case of the

Dear R Users, Is anyone aware of a package which calculates the Yule Kendall resistant (to errors,outliers) measure of skewness ? An easy calculation to perform, but was just wondering if a package exists (as the contents of that package would probably include other cool things I would also be interested). Thanks, Tolga Generally, this communication is for informational purposes only and it

> On Sep 13, 2019, at 10:45 AM, Chandler Carruth <chandlerc at gmail.com> wrote: > > On Fri, Sep 13, 2019 at 1:33 AM David Zarzycki via llvm-dev <llvm-dev at lists.llvm.org <mailto:llvm-dev at lists.llvm.org>> wrote: > Hi Chandler, > > The data-invariant feature sounds great but what about the general case? When performance tuning code, people sometimes need

I’m struggling to find cases where __builtin_unpredictable() works at all. Even if we ignore cmp/br into switch conversion, it still doesn’t work: int test_cmov(int left, int right, int *alt) { return __builtin_unpredictable(left < right) ? *alt : 999; } Should generate: test_cmov: movl $999, %eax cmpl %esi, %edi cmovll (%rdx), %eax retq But currently generates: test_cmov: movl $999,

Hi Chandler, I feel like this conversation has come full circle. So to ask again: how does one force CMOV to be emitted? You suggested “__builtin_unpredictable()” but that gets lost in various optimization passes. Given other architectures have CMOV like instructions, and given the usefulness of the instruction for performance tuning, it seems like a direct intrinsic would be best. What am I

On Tue, Nov 3, 2015 at 3:15 PM, Daniel Berlin <dberlin at dberlin.org> wrote: > > > On Tue, Nov 3, 2015 at 12:29 PM, Richard Diamond < > wichard at vitalitystudios.com> wrote: > >> >> >> On Mon, Nov 2, 2015 at 9:16 PM, Daniel Berlin <dberlin at dberlin.org> >> wrote: >> >>> I'm very unclear and why you think a generic

Dear All I am kinda frustrated with the lack of help some developers are dispensing to this problem (read ignoring it). A few days ago (december 22th) I reported a problem with ''tc filter match mark'' against 2.4 kernel series and got no answers that could lead me to any solution. This is a steady problem that occurs in kernel 2.4 series and match mark (at least for me). I

Hello, in addition to my last thread about a new config option to make strict-kex mandatory, I also wonder if a new mechanism for ciphers/macs can be introduced and is reliable by simple both sides using it. So there could be a Chacha20-Poly1305v2 at openssh.com which uses AD data to chain the messages together, so it will be resistant against terrapin even without the strict-kex. Consequently

Hi there, > So there could be a Chacha20-Poly1305v2 at openssh.com which uses AD data to chain the > messages together, so it will be resistant against terrapin even without the strict-kex. > > Consequently the hmac-etmv2 at openssh.com mode could be deviced in a similar manner, to > also include the transcript hash or similar things. This would still require both, client and