Commit message (Collapse) | Author | Age | ||
---|---|---|---|---|
... | ||||
* | Add example usage | Jason Gross | 2017-09-27 | |
| | ||||
* | Update etc scripts to include governor | Jason Gross | 2017-09-27 | |
| | | | | | | | | It needs to be in performance, not powersave, to work well on my machine. While we're at it, also have the scripts print usage if you pass no arguments, rather than giving an error message about $1 being unset. | |||
* | Add curve25519-donna-c64 to etc/third_party | Jason Gross | 2017-09-27 | |
| | ||||
* | Add missing file | Jason Gross | 2017-09-21 | |
| | ||||
* | Add femul,fesqure for C32 | Jason Gross | 2017-09-21 | |
| | | | | | 32-bit ladderstep takes way too long (at least on Coq 8.6), so we don't add it yet | |||
* | Update constants files | Jason Gross | 2017-09-21 | |
| | ||||
* | Add extract_Expr | Jason Gross | 2017-09-21 | |
| | ||||
* | Split off tactics in IntegrationTestDisplayCommon | Jason Gross | 2017-09-21 | |
| | ||||
* | Check if /sys/devices/system/cpu/intel_pstate/no_turbo exists before ↵ | Jason Gross | 2017-09-16 | |
| | | | | searching it | |||
* | Drop test target on travis | Jason Gross | 2017-09-16 | |
| | | | | It gives "illegal instruction" on p256 | |||
* | Test `make bench` on travis. | Jason Gross | 2017-09-15 | |
| | ||||
* | Don't inline measure_helper | Jason Gross | 2017-09-15 | |
| | | | | It breaks the ability to use bp in inline asm | |||
* | Allow bp in inline asm by tweaking measure | Jason Gross | 2017-09-15 | |
| | ||||
* | Merge remote-tracking branch 'origin/master' | Jason Gross | 2017-09-15 | |
|\ | ||||
| * | Use imul (not sure if it's better...) | Jason Gross | 2017-09-12 | |
| | | | | | | | | | | ... but that's what shows up on supercop: https://github.com/floodyberry/supercop/blob/master/crypto_scalarmult/curve25519/amd64-51/fe25519_mul.s#L164 | |||
* | | Use imul (not sure if it's better...) | Jason Gross | 2017-09-12 | |
|/ | ||||
* | make bench | Jason Gross | 2017-09-12 | |
| | ||||
* | Preserved rbp, not rsp | Jason Gross | 2017-09-12 | |
| | | | | Apparently gcc accessess memory via rbp, according to objdump -D | |||
* | Don't claim to clobber rbp | Jason Gross | 2017-09-12 | |
| | | | | gcc complains if we claim to clobber it | |||
* | Be better about asm syntax dialects | Jason Gross | 2017-09-12 | |
| | | | | | | | With some help from stackoverflow, https://stackoverflow.com/questions/46186592/how-do-i-refer-to-literal-registers-in-gcc-inline-assembly-in-att-syntax and https://stackoverflow.com/questions/46185788/how-can-i-pass-an-immediate-value-to-shr-in-assembly-in-intel-syntax | |||
* | Update register allocation more | Jason Gross | 2017-09-12 | |
| | | | | | Switch over to intel syntax, because I can't figure out how to name registers in AT&T / GAS. | |||
* | Update reg alloc | Jason Gross | 2017-09-12 | |
| | ||||
* | Update extract-function.sh for inline asm | Jason Gross | 2017-09-12 | |
| | | | | | It needs to use printf rather than echo to not interpret \n, \t, and it needs to not insert { and } everywhere. | |||
* | Add reg alloc to root | Jason Gross | 2017-09-12 | |
| | ||||
* | Fix assembly | Jason Gross | 2017-09-12 | |
| | ||||
* | WIP on reg alloc for asm output | Jason Gross | 2017-09-11 | |
| | ||||
* | fsatz tests: 1/(1/x) = x | Andres Erbsen | 2017-09-11 | |
| | ||||
* | Fix more commented out alternatives | Jason Gross | 2017-09-05 | |
| | ||||
* | Fix commented out alternate version in ladderstep | Jason Gross | 2017-09-05 | |
| | ||||
* | Fix (hopefully) overlap in reg | Jason Gross | 2017-09-05 | |
| | ||||
* | Fix schedule to be in correct order | Jason Gross | 2017-09-05 | |
| | ||||
* | WIP | Jason Gross | 2017-09-05 | |
| | ||||
* | Fix schedule | Jason Gross | 2017-09-05 | |
| | ||||
* | WIP on reg alloc | Jason Gross | 2017-09-04 | |
| | ||||
* | chmod +x | Jason Gross | 2017-09-04 | |
| | ||||
* | Add display files (temporary, kind-of) | Jason Gross | 2017-09-04 | |
| | ||||
* | WIP Update compile with registers | Jason Gross | 2017-09-04 | |
| | ||||
* | More WIP on register allocation | Jason Gross | 2017-08-17 | |
| | | | | | | | | | | The current allocation is terrible, probably because we are currently requiring that all instructions output to registers. My current guess at a decent thing to do is to make a pass, after register allocation, and eliminate all registers that simply get stored to memory, replacing the relevant instructions with the memory-using versions. Then we can re-register allocate, ignoring values that go straight to memory. | |||
* | Handle most of register allocation | Jason Gross | 2017-08-14 | |
| | ||||
* | Handle equality in parsing | Jason Gross | 2017-08-14 | |
| | ||||
* | Fixup header and footer | Jason Gross | 2017-08-14 | |
| | ||||
* | Update scheduler to know about implicit mulx arg | Jason Gross | 2017-08-14 | |
| | | | | | Now it prefers putting together mulx with the same implicit arg (approximated as the same variable with the lower number). | |||
* | Use a more realistic processor model | Jason Gross | 2017-08-13 | |
| | ||||
* | Revert "Revert "Subset compiler differently"" | Jason Gross | 2017-08-13 | |
| | | | | This reverts commit c8c82f2edf416b667ba487dfa2cff0795d37cbe6. | |||
* | Revert "Subset compiler differently" | Jason Gross | 2017-08-13 | |
| | | | | This reverts commit 7038444e6515ec51a3b6d1cdfe972664d3f16c81. | |||
* | Subset compiler differently | Jason Gross | 2017-08-13 | |
| | ||||
* | Add heuristic search | Jason Gross | 2017-08-13 | |
| | ||||
* | Add decidable equality with nil | Jason Gross | 2017-08-13 | |
| | ||||
* | Update the graph maker | Jason Gross | 2017-08-13 | |
| | ||||
* | Update exhaustive search compiler | Jason Gross | 2017-08-13 | |
| |