Mahesh Madhav
Mahesh Madhav
Many people download graphs from the Stanford SNAP database. These are edgelist files (.el) however they contain some comments as a header. Manually spicing out the first four lines of...
This is an easy one to fix for anyone wanting to contribute. (I can get to it later if nobody else does). If one supplies `-su` as a parameter alongside...
Hi, I have noticed that opus_compare is not a reflexive function. That is, the result of opus_compare is dependent on the order of the arguments. Is this expected? ``` $...
Hi @bcoconni and friends, I’m a CPU architect at Ampere Computing where I do performance analysis and workload characterization. I also serve on the [SPEC CPU](https://spec.org/cpu2017/) committee, searching for benchmarks...
Hi folks, I'm looking to test HighwayHash on our new Ampere Altra to ensure performance. I'm having some trouble running. I was wondering if you could share your invocations? I...
flac is being considered for the next version of the SPEC CPU benchmark suite, currently dubbed [CPUv8](https://spec.org/cpuv8/). As such, we are interested in the generic (non-intrinsic) code paths that can...
Hello friends, I’m a CPU architect at Ampere Computing where I do performance analysis and workload characterization. I also serve on the [SPEC CPU](https://spec.org/cpu2017/) committee, working on benchmarks for the...
This patch reduces read syscalls by 96%. `strace` showed that we were doing small incremental reads of large files: ``` openat(AT_FDCWD, "stratixiv_arch.timing.xml", O_RDONLY) = 5 read(5, "\n \n \n
### Description For the generic/CPU path, we noticed a super hot loop executing a floating point divide, where the divisor is a loop invariant and can be replaced by a...
`std::endl` flushes the output stream which invokes the `write` syscall. Replacing this with `"\n"` causes the writes to be buffered until they fill the entire memory page or buffer, and...