bwlp/qemu.git/target/ppc/translate, branch master

target/ppc: Use gvec to decode XVTSTDC[DS]P

2022-10-28T16:15:22+00:00

Used gvec to translate XVTSTDCSP and XVTSTDCDP. xvtstdcsp: rept loop imm master version prev version current version 25 4000 0 0,206200 0,040730 (-80.2%) 0,040740 (-80.2%) 25 4000 1 0,205120 0,053650 (-73.8%) 0,053510 (-73.9%) 25 4000 3 0,206160 0,058630 (-71.6%) 0,058570 (-71.6%) 25 4000 51 0,217110 0,191490 (-11.8%) 0,192320 (-11.4%) 25 4000 127 0,206160 0,191490 (-7.1%) 0,192640 (-6.6%) 8000 12 0 1,234719 0,418833 (-66.1%) 0,386365 (-68.7%) 8000 12 1 1,232417 1,435979 (+16.5%) 1,462792 (+18.7%) 8000 12 3 1,232760 1,766073 (+43.3%) 1,743990 (+41.5%) 8000 12 51 1,239281 1,319562 (+6.5%) 1,423479 (+14.9%) 8000 12 127 1,231708 1,315760 (+6.8%) 1,426667 (+15.8%) xvtstdcdp: rept loop imm master version prev version current version 25 4000 0 0,159930 0,040830 (-74.5%) 0,040610 (-74.6%) 25 4000 1 0,160640 0,053670 (-66.6%) 0,053480 (-66.7%) 25 4000 3 0,160020 0,063030 (-60.6%) 0,062960 (-60.7%) 25 4000 51 0,160410 0,128620 (-19.8%) 0,127470 (-20.5%) 25 4000 127 0,160330 0,127670 (-20.4%) 0,128690 (-19.7%) 8000 12 0 1,190365 0,422146 (-64.5%) 0,388417 (-67.4%) 8000 12 1 1,191292 1,445312 (+21.3%) 1,428698 (+19.9%) 8000 12 3 1,188687 1,980656 (+66.6%) 1,975354 (+66.2%) 8000 12 51 1,191250 1,264500 (+6.1%) 1,355083 (+13.8%) 8000 12 127 1,197313 1,266729 (+5.8%) 1,349156 (+12.7%) Overall, these instructions are the hardest ones to measure performance as the gvec implementation is affected by the immediate. Above there are 5 different scenarios when it comes to immediate and 2 when it comes to rept/loop combination. The immediates scenarios are: all bits are 0 therefore the target register should just be changed to 0, with 1 bit set, with 2 bits set in a combination the new implementation can deal with using gvec, 4 bits set and the new implementation can't deal with it using gvec and all bits set. The rept/loop scenarios are high loop and low rept (so it should spend more time executing it than translating it) and high rept low loop (so it should spend more time translating it than executing this code). These comparisons are between the upstream version, a previous similar implementation and a one with a cleaner code(this one). For a comparison with o previous different implementation: <20221010191356.83659-13-lucas.araujo@eldorado.org.br> Signed-off-by: Lucas Mateus Castro (alqotel) Reviewed-by: Richard Henderson Message-Id: <20221019125040.48028-13-lucas.araujo@eldorado.org.br> Signed-off-by: Daniel Henrique Barboza

target/ppc: Moved XSTSTDC[QDS]P to decodetree

2022-10-28T16:15:22+00:00

Moved XSTSTDCSP, XSTSTDCDP and XSTSTDCQP to decodetree and moved some of its decoding away from the helper as previously the DCMX, XB and BF were calculated in the helper with the help of cpu_env, now that part was moved to the decodetree with the rest. xvtstdcsp: rept loop master patch 8 12500 1,85393600 1,94683600 (+5.0%) 25 4000 1,78779800 1,92479000 (+7.7%) 100 1000 2,12775000 2,28895500 (+7.6%) 500 200 2,99655300 3,23102900 (+7.8%) 2500 40 6,89082200 7,44827500 (+8.1%) 8000 12 17,50585500 18,95152100 (+8.3%) xvtstdcdp: rept loop master patch 8 12500 1,39043100 1,33539800 (-4.0%) 25 4000 1,35731800 1,37347800 (+1.2%) 100 1000 1,51514800 1,56053000 (+3.0%) 500 200 2,21014400 2,47906000 (+12.2%) 2500 40 5,39488200 6,68766700 (+24.0%) 8000 12 13,98623900 18,17661900 (+30.0%) xvtstdcdp: rept loop master patch 8 12500 1,35123800 1,34455800 (-0.5%) 25 4000 1,36441200 1,36759600 (+0.2%) 100 1000 1,49763500 1,54138400 (+2.9%) 500 200 2,19020200 2,46196400 (+12.4%) 2500 40 5,39265700 6,68147900 (+23.9%) 8000 12 14,04163600 18,19669600 (+29.6%) As some values are now decoded outside the helper and passed to it as an argument the number of arguments of the helper increased, the number of TCGop needed to load the arguments increased. I suspect that's why the slow-down in the tests with a high REPT but low LOOP. Signed-off-by: Lucas Mateus Castro (alqotel) Reviewed-by: Richard Henderson Message-Id: <20221019125040.48028-12-lucas.araujo@eldorado.org.br> Signed-off-by: Daniel Henrique Barboza

target/ppc: Moved XVTSTDC[DS]P to decodetree

2022-10-28T16:15:22+00:00

Moved XVTSTDCSP and XVTSTDCDP to decodetree an restructured the helper to be simpler and do all decoding in the decodetree (so XB, XT and DCMX are all calculated outside the helper). Obs: The tests in this one are slightly different, these are the sum of these instructions with all possible immediate and those instructions are repeated 10 times. xvtstdcsp: rept loop master patch 8 12500 2,76402100 2,70699100 (-2.1%) 25 4000 2,64867100 2,67884100 (+1.1%) 100 1000 2,73806300 2,78701000 (+1.8%) 500 200 3,44666500 3,61027600 (+4.7%) 2500 40 5,85790200 6,47475500 (+10.5%) 8000 12 15,22102100 17,46062900 (+14.7%) xvtstdcdp: rept loop master patch 8 12500 2,11818000 1,61065300 (-24.0%) 25 4000 2,04573400 1,60132200 (-21.7%) 100 1000 2,13834100 1,69988100 (-20.5%) 500 200 2,73977000 2,48631700 (-9.3%) 2500 40 5,05067000 5,25914100 (+4.1%) 8000 12 14,60507800 15,93704900 (+9.1%) Signed-off-by: Lucas Mateus Castro (alqotel) Reviewed-by: Richard Henderson Message-Id: <20221019125040.48028-11-lucas.araujo@eldorado.org.br> Signed-off-by: Daniel Henrique Barboza

target/ppc: Use gvec to decode XVCPSGN[SD]P

2022-10-28T16:15:22+00:00

Moved XVCPSGNSP and XVCPSGNDP to decodetree and used gvec to translate them. xvcpsgnsp: rept loop master patch 8 12500 0,00561400 0,00537900 (-4.2%) 25 4000 0,00562100 0,00400000 (-28.8%) 100 1000 0,00696900 0,00416300 (-40.3%) 500 200 0,02211900 0,00840700 (-62.0%) 2500 40 0,09328600 0,02728300 (-70.8%) 8000 12 0,27295300 0,06867800 (-74.8%) xvcpsgndp: rept loop master patch 8 12500 0,00556300 0,00584200 (+5.0%) 25 4000 0,00482700 0,00431700 (-10.6%) 100 1000 0,00585800 0,00464400 (-20.7%) 500 200 0,01565300 0,00839700 (-46.4%) 2500 40 0,05766500 0,02430600 (-57.8%) 8000 12 0,19875300 0,07947100 (-60.0%) Like the previous instructions there seemed to be a improvement on translation time. Signed-off-by: Lucas Mateus Castro (alqotel) Reviewed-by: Richard Henderson Message-Id: <20221019125040.48028-10-lucas.araujo@eldorado.org.br> Signed-off-by: Daniel Henrique Barboza

target/ppc: Use gvec to decode XV[N]ABS[DS]P/XVNEG[DS]P

2022-10-28T16:15:22+00:00

Moved XVABSSP, XVABSDP, XVNABSSP,XVNABSDP, XVNEGSP and XVNEGDP to decodetree and used gvec to translate them. xvabssp: rept loop master patch 8 12500 0,00477900 0,00476000 (-0.4%) 25 4000 0,00442800 0,00353300 (-20.2%) 100 1000 0,00478700 0,00366100 (-23.5%) 500 200 0,00973200 0,00649400 (-33.3%) 2500 40 0,03165200 0,02226700 (-29.7%) 8000 12 0,09315900 0,06674900 (-28.3%) xvabsdp: rept loop master patch 8 12500 0,00475000 0,00474400 (-0.1%) 25 4000 0,00355600 0,00367500 (+3.3%) 100 1000 0,00444200 0,00366000 (-17.6%) 500 200 0,00942700 0,00732400 (-22.3%) 2500 40 0,02990000 0,02308500 (-22.8%) 8000 12 0,08770300 0,06683800 (-23.8%) xvnabssp: rept loop master patch 8 12500 0,00494500 0,00492900 (-0.3%) 25 4000 0,00397700 0,00338600 (-14.9%) 100 1000 0,00421400 0,00353500 (-16.1%) 500 200 0,01048000 0,00707100 (-32.5%) 2500 40 0,03251500 0,02238300 (-31.2%) 8000 12 0,08889100 0,06469800 (-27.2%) xvnabsdp: rept loop master patch 8 12500 0,00511000 0,00492700 (-3.6%) 25 4000 0,00398800 0,00381500 (-4.3%) 100 1000 0,00390500 0,00365900 (-6.3%) 500 200 0,00924800 0,00784600 (-15.2%) 2500 40 0,03138900 0,02391600 (-23.8%) 8000 12 0,09654200 0,05684600 (-41.1%) xvnegsp: rept loop master patch 8 12500 0,00493900 0,00452800 (-8.3%) 25 4000 0,00369100 0,00366800 (-0.6%) 100 1000 0,00371100 0,00380000 (+2.4%) 500 200 0,00991100 0,00652300 (-34.2%) 2500 40 0,03025800 0,02422300 (-19.9%) 8000 12 0,09251100 0,06457600 (-30.2%) xvnegdp: rept loop master patch 8 12500 0,00474900 0,00454400 (-4.3%) 25 4000 0,00353100 0,00325600 (-7.8%) 100 1000 0,00398600 0,00366800 (-8.0%) 500 200 0,01032300 0,00702400 (-32.0%) 2500 40 0,03125000 0,02422400 (-22.5%) 8000 12 0,09475100 0,06173000 (-34.9%) This one to me seemed the opposite of the previous instructions, as it looks like there was an improvement in the translation time (itself not a surprise as operations were done twice before so there was the need to translate twice as many TCGop) Signed-off-by: Lucas Mateus Castro (alqotel) Reviewed-by: Richard Henderson Message-Id: <20221019125040.48028-9-lucas.araujo@eldorado.org.br> Signed-off-by: Daniel Henrique Barboza

target/ppc: Move VABSDU[BHW] to decodetree and use gvec

2022-10-28T16:15:22+00:00

Moved VABSDUB, VABSDUH and VABSDUW to decodetree and use gvec to translate them. vabsdub: rept loop master patch 8 12500 0,03601600 0,00688500 (-80.9%) 25 4000 0,03651000 0,00532100 (-85.4%) 100 1000 0,03666900 0,00595300 (-83.8%) 500 200 0,04305800 0,01244600 (-71.1%) 2500 40 0,06893300 0,04273700 (-38.0%) 8000 12 0,14633200 0,12660300 (-13.5%) vabsduh: rept loop master patch 8 12500 0,02172400 0,00687500 (-68.4%) 25 4000 0,02154100 0,00531500 (-75.3%) 100 1000 0,02235400 0,00596300 (-73.3%) 500 200 0,02827500 0,01245100 (-56.0%) 2500 40 0,05638400 0,04285500 (-24.0%) 8000 12 0,13166000 0,12641400 (-4.0%) vabsduw: rept loop master patch 8 12500 0,01646400 0,00688300 (-58.2%) 25 4000 0,01454500 0,00475500 (-67.3%) 100 1000 0,01545800 0,00511800 (-66.9%) 500 200 0,02168200 0,01114300 (-48.6%) 2500 40 0,04571300 0,04138800 (-9.5%) 8000 12 0,12209500 0,12178500 (-0.3%) Same as VADDCUW and VSUBCUW, overall performance gain but it uses more TCGop (4 before the patch, 6 after). Signed-off-by: Lucas Mateus Castro (alqotel) Reviewed-by: Richard Henderson Message-Id: <20221019125040.48028-8-lucas.araujo@eldorado.org.br> Signed-off-by: Daniel Henrique Barboza

target/ppc: Move VAVG[SU][BHW] to decodetree and use gvec

2022-10-28T16:15:22+00:00

Moved the instructions VAVGUB, VAVGUH, VAVGUW, VAVGSB, VAVGSH, VAVGSW, to decodetree and use gvec with them. For these one the right shift had to be made before the sum as to avoid an overflow, so add 1 at the end if any of the entries had 1 in its LSB as to replicate the "+ 1" before the shift described by the ISA. vavgub: rept loop master patch 8 12500 0,02616600 0,00754200 (-71.2%) 25 4000 0,02530000 0,00637700 (-74.8%) 100 1000 0,02604600 0,00790100 (-69.7%) 500 200 0,03189300 0,01838400 (-42.4%) 2500 40 0,06006900 0,06851000 (+14.1%) 8000 12 0,13941000 0,20548500 (+47.4%) vavguh: rept loop master patch 8 12500 0,01818200 0,00780600 (-57.1%) 25 4000 0,01789300 0,00641600 (-64.1%) 100 1000 0,01899100 0,00787200 (-58.5%) 500 200 0,02527200 0,01828400 (-27.7%) 2500 40 0,05361800 0,06773000 (+26.3%) 8000 12 0,12886600 0,20291400 (+57.5%) vavguw: rept loop master patch 8 12500 0,01423100 0,00776600 (-45.4%) 25 4000 0,01780800 0,00638600 (-64.1%) 100 1000 0,02085500 0,00787000 (-62.3%) 500 200 0,02737100 0,01828800 (-33.2%) 2500 40 0,05572600 0,06774200 (+21.6%) 8000 12 0,13101700 0,20311600 (+55.0%) vavgsb: rept loop master patch 8 12500 0,03006000 0,00788600 (-73.8%) 25 4000 0,02882200 0,00637800 (-77.9%) 100 1000 0,02958000 0,00791400 (-73.2%) 500 200 0,03548800 0,01860400 (-47.6%) 2500 40 0,06360000 0,06850800 (+7.7%) 8000 12 0,13816500 0,20550300 (+48.7%) vavgsh: rept loop master patch 8 12500 0,01965900 0,00776600 (-60.5%) 25 4000 0,01875400 0,00638700 (-65.9%) 100 1000 0,01952200 0,00786900 (-59.7%) 500 200 0,02562000 0,01760300 (-31.3%) 2500 40 0,05384300 0,06742800 (+25.2%) 8000 12 0,13240800 0,20330000 (+53.5%) vavgsw: rept loop master patch 8 12500 0,01407700 0,00775600 (-44.9%) 25 4000 0,01762300 0,00640000 (-63.7%) 100 1000 0,02046500 0,00788500 (-61.5%) 500 200 0,02745600 0,01843000 (-32.9%) 2500 40 0,05375500 0,06820500 (+26.9%) 8000 12 0,13068300 0,20304900 (+55.4%) These results to me seems to indicate that with gvec the results have a slower translation but faster execution. Signed-off-by: Lucas Mateus Castro (alqotel) Reviewed-by: Richard Henderson Message-Id: <20221019125040.48028-7-lucas.araujo@eldorado.org.br> Signed-off-by: Daniel Henrique Barboza

target/ppc: Move VPRTYB[WDQ] to decodetree and use gvec

2022-10-28T16:15:22+00:00

Moved VPRTYBW and VPRTYBD to use gvec and both of them and VPRTYBQ to decodetree. VPRTYBW and VPRTYBD now also use .fni4 and .fni8, respectively. vprtybw: rept loop master patch 8 12500 0,01198900 0,00703100 (-41.4%) 25 4000 0,01070100 0,00571400 (-46.6%) 100 1000 0,01123300 0,00678200 (-39.6%) 500 200 0,01601500 0,01535600 (-4.1%) 2500 40 0,03872900 0,05562100 (43.6%) 8000 12 0,10047000 0,16643000 (65.7%) vprtybd: rept loop master patch 8 12500 0,00757700 0,00788100 (4.0%) 25 4000 0,00652500 0,00669600 (2.6%) 100 1000 0,00714400 0,00825400 (15.5%) 500 200 0,01211000 0,01903700 (57.2%) 2500 40 0,03483800 0,07021200 (101.5%) 8000 12 0,09591800 0,21036200 (119.3%) vprtybq: rept loop master patch 8 12500 0,00675600 0,00667200 (-1.2%) 25 4000 0,00619400 0,00643200 (3.8%) 100 1000 0,00707100 0,00751100 (6.2%) 500 200 0,01199300 0,01342000 (11.9%) 2500 40 0,03490900 0,04092900 (17.2%) 8000 12 0,09588200 0,11465100 (19.6%) I wasn't expecting such a performance lost in both VPRTYBD and VPRTYBQ, I'm not sure if it's worth to move those instructions. Comparing the assembly of the helper with the TCGop they are pretty similar, so I'm not sure why vprtybd took so much more time. Signed-off-by: Lucas Mateus Castro (alqotel) Reviewed-by: Richard Henderson Message-Id: <20221019125040.48028-6-lucas.araujo@eldorado.org.br> Signed-off-by: Daniel Henrique Barboza

target/ppc: Move VNEG[WD] to decodtree and use gvec

2022-10-28T16:15:22+00:00

Moved the instructions VNEGW and VNEGD to decodetree and used gvec to decode it. vnegw: rept loop master patch 8 12500 0,01053200 0,00548400 (-47.9%) 25 4000 0,01030500 0,00390000 (-62.2%) 100 1000 0,01096300 0,00395400 (-63.9%) 500 200 0,01472000 0,00712300 (-51.6%) 2500 40 0,03809000 0,02147700 (-43.6%) 8000 12 0,09957100 0,06202100 (-37.7%) vnegd: rept loop master patch 8 12500 0,00594600 0,00543800 (-8.5%) 25 4000 0,00575200 0,00396400 (-31.1%) 100 1000 0,00676100 0,00394800 (-41.6%) 500 200 0,01149300 0,00709400 (-38.3%) 2500 40 0,03441500 0,02169600 (-37.0%) 8000 12 0,09516900 0,06337000 (-33.4%) Signed-off-by: Lucas Mateus Castro (alqotel) Reviewed-by: Richard Henderson Message-Id: <20221019125040.48028-5-lucas.araujo@eldorado.org.br> Signed-off-by: Daniel Henrique Barboza

target/ppc: Move V(ADD|SUB)CUW to decodetree and use gvec

2022-10-28T16:15:22+00:00

This patch moves VADDCUW and VSUBCUW to decodtree with gvec using an implementation based on the helper, with the main difference being changing the -1 (aka all bits set to 1) result returned by cmp when true to +1. It also implemented a .fni4 version of those instructions and dropped the helper. vaddcuw: rept loop master patch 8 12500 0,01008200 0,00612400 (-39.3%) 25 4000 0,01091500 0,00471600 (-56.8%) 100 1000 0,01332500 0,00593700 (-55.4%) 500 200 0,01998500 0,01275700 (-36.2%) 2500 40 0,04704300 0,04364300 (-7.2%) 8000 12 0,10748200 0,11241000 (+4.6%) vsubcuw: rept loop master patch 8 12500 0,01226200 0,00571600 (-53.4%) 25 4000 0,01493500 0,00462100 (-69.1%) 100 1000 0,01522700 0,00455100 (-70.1%) 500 200 0,02384600 0,01133500 (-52.5%) 2500 40 0,04935200 0,03178100 (-35.6%) 8000 12 0,09039900 0,09440600 (+4.4%) Overall there was a gain in performance, but the TCGop code was still slightly bigger in the new version (it went from 4 to 5). Signed-off-by: Lucas Mateus Castro (alqotel) Reviewed-by: Richard Henderson Message-Id: <20221019125040.48028-4-lucas.araujo@eldorado.org.br> Signed-off-by: Daniel Henrique Barboza