Thank you for your information! I will probably try to port the C version to Go and see if the C tricks can be applied to the Go code. I am reluctant to work with cgo, to be honest.
BTW, the C implementation does not come from OpenSSL. I will check its licence and see if I can share publicly. On Sunday, August 28, 2016 at 3:53:46 AM UTC+8, agl wrote: > > On Saturday, August 27, 2016 at 9:43:23 AM UTC-7, mura wrote: >> >> We have some legacy systems which use 3DES cryptos extensively and we are >> attempting to re-write some parts of it in Go recently. >> As an initial experiment, we found that the system overall is about 2x~3x >> slower than the original C version. >> Although the culprit can not be completely singled out yet, the Go pprof >> tool has shown that about half of the execution time is spent on the DES >> operation (and the feistel function, specifically). >> > > It's likely that you were previously using OpenSSL's DES code if it was > written in C. I don't think that OpenSSL has ARM assembly for DES, but it > does have a lot of a macro-ridden, hand-unrolled code for implementing DES. > It's not a different algorithm, but the Go code is written to be concise > rather than to eek out performance. > > I'm sure that the Go code could be made faster (at the cost of code size > and clarity) but if 3DES is a critical hot spot for you, the quickest > solution may be to wrap OpenSSL's DES function in cgo. > > > Cheers > > AGL > -- You received this message because you are subscribed to the Google Groups "golang-nuts" group. To unsubscribe from this group and stop receiving emails from it, send an email to golang-nuts+unsubscr...@googlegroups.com. For more options, visit https://groups.google.com/d/optout.
