01:29
itamarst has joined #pypy
03:47
itamarst has quit [Quit: Connection closed for inactivity]
06:29
BarrensZeppelin has joined #pypy
06:31
<
BarrensZeppelin >
cfbolz The blog post on the known bits abstract domain was a nice read. 🙌
07:13
<
cfbolz >
BarrensZeppelin: thank you! I'm glad someone's reading it 😅
07:56
<
BarrensZeppelin >
🙂
07:56
<
BarrensZeppelin >
Do you have an intuition about how the new optimizations will affect PyPy performance?
07:56
<
BarrensZeppelin >
Of course it's application specific, but if the interpreter code itself uses lots of bitwise operations (for some reason / maybe for GC?), it could yield nice speedups for most applications.
08:03
BarrensZeppelin has quit [Quit: Client closed]
08:05
<
xorAxAx >
yeah, nobody except RSqueak uses that seriously, but dont tell anybody :)
08:11
<
cfbolz >
BarrensZeppelin: it doesn't really help most python applications, unfortunately
09:08
<
mjacob >
cfbolz: i didn't know about pydrofoil. nice!
09:08
<
mjacob >
cfbolz: do you know who uses it?
09:14
<
cfbolz >
mjacob: probably no one, to a first order of approximation 😂
09:14
<
cfbolz >
mjacob: it's my pet project, I have some more features planned that could make it more useful (like python scriptability)
09:15
<
cfbolz >
it's fairly far though, can boot linux on both arm64 and risc-v64
09:15
<
mjacob >
more useful for those who are disappointed by the 18x speedup?
09:15
<
cfbolz >
mjacob: it's 18x compared to the extremely slow sail baseline interpreter. qemu is way faster
09:16
<
nikolar >
cfbolz: sorry, i wasn't here earlier, what exactly is your project
09:16
<
mjacob >
do you know how it compares to the approach chosen by pydgin?
09:18
<
cfbolz >
mjacob: pydgin is probably faster, but it's a user mode only emulator
09:19
<
cfbolz >
mjacob: pydrofoil supports the various exception levels, and all the (in arm64 extremely complex!) details of address translation
09:20
<
cfbolz >
I have a todo to also add support and tests for the hypervisor extension of risc-v
09:22
<
nikolar >
cool project cfbolz
09:38
Dejan has joined #pypy
09:39
<
larstiq_ >
ooh, I have two blogposts to read
09:40
<
cfbolz >
hehe, that's a very indirect way to add to people's todo list
09:43
<
larstiq_ >
well, I do enjoy your posts so it's not a todo in that sense :)
09:44
<
larstiq_ >
it's just that I'm currently working so will take half a day before I can
09:58
Diggsey has joined #pypy
11:15
jcea has joined #pypy
13:04
Dejan has quit [Quit: Leaving]
14:18
<
korvo >
cfbolz: I haven't dug into the source; do you do knownbits for bigints too, or just for int32/int64?
14:19
<
korvo >
knownbits for bigints could possibly speed up algos which use bigints as bitsets.
14:21
<
cfbolz >
korvo: just word-sized ints
14:22
<
korvo >
cfbolz: Maybe it's time to revive tagged pointers?
14:22
<
korvo >
Also I should add pydrofoil to rpypkgs.
14:22
<
cfbolz >
tagged pointers, unlikely. it's a mess
14:42
<
cfbolz >
korvo: just as a warning, pydrofoil-arm is a real beast to build
14:44
<
korvo >
cfbolz: Understandable. I don't think I have ARM machines for that.
14:44
<
cfbolz >
no, pydrofoil-arm is version where arm is the guest CPU
14:44
<
cfbolz >
it's also a beast to build on an x86 host
14:48
<
korvo >
What's the constraint? I have a chonky second-hand workstation with 150 GiB of RAM, and Nix should make dependency management easy.
14:50
<
cfbolz >
if you have that much RAM it's ok. will just take a while (about 4x longer than a pypy build)
14:56
itamarst has joined #pypy
15:06
<
korvo >
Noted, thanks. I appreciate the warning.
15:36
jcea has quit [Ping timeout: 252 seconds]
15:38
otisolsen70 has joined #pypy
19:37
Diggsey has quit [Quit: Connection closed for inactivity]
19:49
otisolsen70 has quit [Read error: Connection reset by peer]
20:25
Dejan has joined #pypy
23:18
jcea has joined #pypy