--- Log opened Thu Oct 18 00:00:40 2018
04:25 -!- ken2812221 [~ken281222@110.50.135.178] has joined #secp256k1
05:41 -!- klot [~klot@188.113.35.204] has joined #secp256k1
05:42 -!- klot [~klot@188.113.35.204] has quit [Remote host closed the connection]
05:42 -!- klot [~klot@188.113.35.204] has joined #secp256k1
05:43 -!- klot [~klot@188.113.35.204] has quit [Remote host closed the connection]
05:44 -!- klot [~klot@188.113.35.204] has joined #secp256k1
05:45 -!- klot [~klot@188.113.35.204] has quit [Remote host closed the connection]
05:45 -!- klot [~klot@188.113.35.204] has joined #secp256k1
12:05 -!- belcher [~belcher@unaffiliated/belcher] has quit [Quit: Leaving]
12:10 -!- belcher [~belcher@unaffiliated/belcher] has joined #secp256k1
12:17 -!- belcher_ [~belcher@unaffiliated/belcher] has joined #secp256k1
12:18 -!- belcher [~belcher@unaffiliated/belcher] has quit [Ping timeout: 260 seconds]
15:29 -!- ddustin [40470882@gateway/web/freenode/ip.64.71.8.130] has joined #secp256k1
15:29 < ddustin> I'm trying to get the library smaller by 40KB. What is taking up all this space that I might be able to remove?
15:33 < ddustin> It seems odd the library would compile to something so huge
15:34 < sipa> there are large precomputed tables
15:37 < ddustin> Is there a way around using them? I wonder how mbed TLS manages to do it
15:37 < sipa> secp256k1 arithmetic can probably be implemented in a kilobyte of code if you don't care about performance
15:38 < sipa> but libsecp256k1 is a highly optimized library
15:38 < ddustin> Ah I'm seeing a constant 'USE_ECMULT_STATIC_PRECOMPUTATION' -- would turning that off disable it?
15:39 < sipa> it would reduce the binary size, not memory usage
15:39 < sipa> (it would instead be created at runtime)
15:40 < ddustin> Makes sense. Would it be created once and saved?
15:41 < andytoshi> to disable it use --enable-ecmult-static-precomputation=no rather than directly messing with that constant (though i guess it'd have the same effect)
15:41 < andytoshi> ddustin: it's created when you call secp256k1_context_create
15:41 < andytoshi> and will live as long as you hold onto the returned context object
15:41 < andytoshi> which should be destroyed with secp256k1_context_destroy when you're done with it
15:42 < andytoshi> as sipa says, you're just trading static memory for runtime/allocated memory here .. i don't believe we expose any way to shrink the precomp table
15:43 < gmaxwell> ddustin: What do you mean specifically by smaller than 40kb?
15:44 < gmaxwell> e.g. in embedded devices memory usage is critical, but size of the const static data is mostly irrelevant.
15:44 < gmaxwell> (a typical hardware wallet has 32kb ram, and 256KB flash)
15:48 < ddustin> gmaxwell: Yeah putting it in the flash probably makes sense
15:50 < ddustin> I'm just trying to get it to build at all at the moment, wrangling multiple build systems each with their own problems. Shrinking the build size will just get me to the next step more easily -- can come back and revisit a more effective solution later
16:05 -!- klot [~klot@188.113.35.204] has quit [Read error: Connection reset by peer]
16:21 -!- lzfernandes [~lzfernand@41.205.52.107] has joined #secp256k1
16:22 -!- ddustin [40470882@gateway/web/freenode/ip.64.71.8.130] has quit [Ping timeout: 256 seconds]
16:26 -!- lzfernandes [~lzfernand@41.205.52.107] has quit [Quit: Leaving]
18:06 -!- ken2812221 [~ken281222@110.50.135.178] has quit [Quit: Leaving]
18:12 -!- ddustin [40470882@gateway/web/freenode/ip.64.71.8.130] has joined #secp256k1
18:12 < ddustin> Hm, would it be possible to use just the ECDH and secp256k1_ec_pubkey_tweak_add parts of the library without the ECMULT precomputations?
18:14 < ddustin> Basically I've realized this chipset has hardware ECC that I should be using -- but I still need to do tweaks and diffie's.
18:37 -!- ddustin [40470882@gateway/web/freenode/ip.64.71.8.130] has quit [Ping timeout: 256 seconds]
18:37 -!- sipa [~pw@gateway/tor-sasl/sipa1024] has quit [Remote host closed the connection]
18:38 -!- sipa [~pw@gateway/tor-sasl/sipa1024] has joined #secp256k1
19:05 -!- lukedashjr [~luke-jr@unaffiliated/luke-jr] has joined #secp256k1
19:06 -!- lukedashjr [~luke-jr@unaffiliated/luke-jr] has quit [Excess Flood]
19:07 -!- lukedashjr [~luke-jr@unaffiliated/luke-jr] has joined #secp256k1
19:09 -!- luke-jr [~luke-jr@unaffiliated/luke-jr] has quit [Ping timeout: 250 seconds]
19:11 -!- lukedashjr is now known as luke-jr
19:39 < andytoshi> ECDH, yes
19:40 < andytoshi> but tweak_add uses the ecumlt precomp
--- Log closed Fri Oct 19 00:00:41 2018