Friday, 2023-06-16

lkclghostmansd[m], lots to go over - i'm still very busy keeping my guest occupied, so i can only be brief, apologies00:21
*** awilfox <awilfox!~awilfox@kelsey.foxkit.us> has joined #libre-soc04:12
programmerjakeas mentioned in https://bugs.libre-soc.org/show_bug.cgi?id=1072#c45 I want to remove Rc=1 from fmvfg since it's unnecessary since fp Rc=1 sets CR1 to some FPSCR bits, yet fmvfg doesn't modify FPSCR at all. lkcl and others, what do you think?04:29
ghostmansd[m]Hi Luke, looks like there's a confusion in 1094. In the last implementation I posted there's no need to even invent Node class, paths kinda form node along with the element. Let's wait till next week then when you have some time.06:49
*** jn <jn!~quassel@user/jn/x-3390946> has quit IRC06:49
*** jn <jn!~quassel@2a0a-a541-c46d-0-20d-b9ff-fe49-15fc.ipv6dyn.netcologne.de> has joined #libre-soc06:54
*** jn <jn!~quassel@2a0a-a541-c46d-0-20d-b9ff-fe49-15fc.ipv6dyn.netcologne.de> has quit IRC06:54
*** jn <jn!~quassel@user/jn/x-3390946> has joined #libre-soc06:54
*** ghostmansd[m] <ghostmansd[m]!~ghostmans@broadband-109-173-83-100.ip.moscow.rt.ru> has quit IRC07:25
*** ghostmansd[m] <ghostmansd[m]!~ghostmans@176.59.57.89> has joined #libre-soc07:25
*** ghostmansd[m] <ghostmansd[m]!~ghostmans@176.59.57.89> has quit IRC07:33
*** ghostmansd[m] <ghostmansd[m]!~ghostmans@176.59.57.89> has joined #libre-soc07:34
*** ghostmansd[m] <ghostmansd[m]!~ghostmans@176.59.57.89> has quit IRC09:24
*** ghostmansd[m] <ghostmansd[m]!~ghostmans@broadband-109-173-83-100.ip.moscow.rt.ru> has joined #libre-soc09:25
*** ghostmansd <ghostmansd!~ghostmans@broadband-109-173-83-100.ip.moscow.rt.ru> has quit IRC09:28
*** ghostmansd[m] <ghostmansd[m]!~ghostmans@broadband-109-173-83-100.ip.moscow.rt.ru> has quit IRC09:56
*** ghostmansd[m] <ghostmansd[m]!~ghostmans@176.59.160.236> has joined #libre-soc09:57
*** sadoon[m] <sadoon[m]!~sadoonsou@2001:470:69fc:105::2:bab8> has quit IRC10:07
*** cesar <cesar!~cesar@2001:470:69fc:105::76c> has quit IRC10:07
*** ghostmansd[m] <ghostmansd[m]!~ghostmans@176.59.160.236> has quit IRC10:09
lkclprogrammerjake, what's it going to look like in the spec?10:11
lkclRc=1 tests FRT and sets CR1, exactly like all other Scalar Floating-Point operations.10:12
programmerjakeprobably either XO extended to include bit 31 or bit 31 just being ignored10:13
lkclwhat does VSX int-to-FP do? that's the behaviour to copy (exactly)10:13
lkcl6.10.3 somewhere...10:14
programmerjakethe vsx equivalent is mtvsrd10:15
programmerjakeit has no Rc field10:15
programmerjakesection 3.3.1810:15
programmerjakenote fmvfg is just copying bits, it is not a int <-> fp conversion10:16
lkclVector Convert with round to nearest From10:16
lkclSigned Word to floating-point format VX-form10:16
lkclvcfsx         VRT,VRB,UIM10:16
lkclahhh10:16
programmerjakevcfsx is not the right one10:17
lkclokaay yes bit-move? then definitely not10:17
lkclgood catch10:17
programmerjakefmvtg can have Rc since that's based on RT which is useful10:18
programmerjakefmvfg, not so much10:18
lkclthere is a *really* good reason why not to do anything and it's because in an OoO system with unified (tagged) in-flight Reservation Stations the entire operation can be elided as a simple "change the tag"10:18
lkcllet me just edit the spec...10:19
programmerjakeyes, that's what the Rc=0 version can be10:19
programmerjakeactually, i'll edit the spec tomorrow (probably)10:19
lkclok10:19
lkclyes please10:20
programmerjakeRc=1 is still useful, since it lets you know if it's +0 and for f64 if the sign bit is set10:20
lkclyes10:20
lkclcan you put "equivalent to mtvsrd" in the spec page as a cross-reference?10:21
lkclok i gotta get back to my guest10:21
programmerjakesince +0 == 0 and sign bit set maps to negative integers10:21
programmerjakek, have fun! though please look at the email about grants since i need more money soon10:22
programmerjakeadded that equivalence note as a todo in https://bugs.libre-soc.org/show_bug.cgi?id=1015#c0 as well as updating a few other todos10:27
*** ghostmansd[m] <ghostmansd[m]!~ghostmans@176.59.42.175> has joined #libre-soc10:29
*** ghostmansd[m] <ghostmansd[m]!~ghostmans@176.59.42.175> has quit IRC10:36
*** ghostmansd[m] <ghostmansd[m]!~ghostmans@broadband-109-173-83-100.ip.moscow.rt.ru> has joined #libre-soc10:36
*** ghostmansd[m] <ghostmansd[m]!~ghostmans@broadband-109-173-83-100.ip.moscow.rt.ru> has quit IRC10:45
*** ghostmansd <ghostmansd!~ghostmans@broadband-109-173-83-100.ip.moscow.rt.ru> has joined #libre-soc11:39
*** ghostmansd <ghostmansd!~ghostmans@broadband-109-173-83-100.ip.moscow.rt.ru> has quit IRC12:00
*** ghostmansd <ghostmansd!~ghostmans@broadband-109-173-83-100.ip.moscow.rt.ru> has joined #libre-soc14:43
lkclghostmansd, ahhh   line 22:     @core.hook(object)15:29
lkclfrickin awesome.  of course object is the base class of everything.15:29
*** ghostmansd[m] <ghostmansd[m]!~ghostmans@176.59.166.18> has joined #libre-soc15:40
*** ghostmansd[m] <ghostmansd[m]!~ghostmans@176.59.166.18> has quit IRC16:09
ghostmansdlkcl, ah, you checked it :-)18:34
ghostmansdyeah, the fact that it goes in class MRO all way down to object is awesome18:34
ghostmansdand the default handler is... object!18:34
*** ghostmansd[m] <ghostmansd[m]!~ghostmans@broadband-109-173-83-100.ip.moscow.rt.ru> has joined #libre-soc20:14
*** ghostmansd <ghostmansd!~ghostmans@broadband-109-173-83-100.ip.moscow.rt.ru> has quit IRC21:03
*** ghostmansd <ghostmansd!~ghostmans@broadband-109-173-83-100.ip.moscow.rt.ru> has joined #libre-soc21:03
*** ghostmansd <ghostmansd!~ghostmans@broadband-109-173-83-100.ip.moscow.rt.ru> has quit IRC21:46
*** ghostmansd <ghostmansd!~ghostmans@broadband-109-173-83-100.ip.moscow.rt.ru> has joined #libre-soc22:47
lkclit is quite mad how much effort goes into something so.... so... stupidly short amount of code23:27
ghostmansd[m]Yeah, I hope at least the benefits will justify it23:42
*** ghostmansd <ghostmansd!~ghostmans@broadband-109-173-83-100.ip.moscow.rt.ru> has quit IRC23:50

Generated by irclog2html.py 2.17.1 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!