Page MenuHome GnuPG

Members

  • This project does not have any members.
  • View All

Watchers

  • This project does not have any watchers.
  • View All

Details

Description

PowerPC related report

Recent Activity

Sun, Jun 12

jukivili closed T6006: Optimize Chacha20 and Poly1305 for PPC P10 LE as Resolved.
Sun, Jun 12, 9:58 PM · patch, ppc, Feature Request, libgcrypt
jukivili added a comment to T6006: Optimize Chacha20 and Poly1305 for PPC P10 LE.

Patch applied to master with small changes.

Sun, Jun 12, 9:58 PM · patch, ppc, Feature Request, libgcrypt

Fri, Jun 3

dannytsen added a comment to T6006: Optimize Chacha20 and Poly1305 for PPC P10 LE.

Thanks @jukivili , Here is the changelog,

Fri, Jun 3, 3:59 PM · patch, ppc, Feature Request, libgcrypt
jukivili added a comment to T6006: Optimize Chacha20 and Poly1305 for PPC P10 LE.

Thanks for updated patch. I'm travelling next week and have time to check it closely only after I'm back. On quick glance, it looks good. What is also needed is the changelog for git commit log.

Fri, Jun 3, 10:30 AM · patch, ppc, Feature Request, libgcrypt

Thu, Jun 2

dannytsen added a comment to T6006: Optimize Chacha20 and Poly1305 for PPC P10 LE.

Thanks @jukivili. I have never thought of interleaving with interger poly1305 operation and that's a good suggestion. Will think about that one.

Thu, Jun 2, 12:59 AM · patch, ppc, Feature Request, libgcrypt

Wed, Jun 1

jukivili added a comment to T6006: Optimize Chacha20 and Poly1305 for PPC P10 LE.

I meant interleaving integer register based 1xPoly1305 with 8xChacha20 as is done for 4xChacha20 in cipher/chacha20-ppc.c (interleaved so that for each 4xChaCha20 processed, 4 blocks of 1xPoly1305 is executed). Quite often microarchitectures have separate execution units for integer registers and vector registers and then it makes sense to interleave integer-poly1305 with vector-chacha20 as algorithms do end up not compete for same execution resources. Interleaving vector-poly1305 and vector-chacha20 is not likely to give performance increase (and likely to run problems with running out of vector registers).

Wed, Jun 1, 5:37 PM · patch, ppc, Feature Request, libgcrypt
dannytsen added a comment to T6006: Optimize Chacha20 and Poly1305 for PPC P10 LE.

HI @jukivili , Thanks for the updates. For f14-f31 registers that was my mistake that did not think floating point will be used. Will correct that. For poly1305, it can be used on ARCH_3.0 so checking use_p10 doesn't seem to be necessary but I can include that as well.

Wed, Jun 1, 4:14 PM · patch, ppc, Feature Request, libgcrypt
gniibe claimed T5980: compilation error libgcrypt 1.10.1.
Wed, Jun 1, 5:08 AM · Testing, backport, ppc, AIX, libgcrypt, Bug Report

May 28 2022

jukivili added a comment to T6006: Optimize Chacha20 and Poly1305 for PPC P10 LE.

Problem is that new assembly is using VSX registers vs14-vs31 which overlap with floating-point registers f14-f31. f14-f31 are ABI callee saved, so those need to be stored and restored.

May 28 2022, 9:04 PM · patch, ppc, Feature Request, libgcrypt
jukivili added a comment to T6006: Optimize Chacha20 and Poly1305 for PPC P10 LE.

Tested patch with small change so that HWF_PPC_ARCH_3_00 is used instead of HWF_PPC_ARCH_3_10. Building bench-slope with "-O3 -flto" makes bug in new implementation visible. Without new implementations bench-slope is ok (testing with QEMU):

$ tests/bench-slope --disable-hwf ppc-arch_3_00 cipher chacha20
Cipher:
 CHACHA20       |  nanosecs/byte   mebibytes/sec   cycles/byte
     STREAM enc |      2.35 ns/B     405.0 MiB/s         - c/B
     STREAM dec |      2.32 ns/B     410.7 MiB/s         - c/B
   POLY1305 enc |      2.46 ns/B     388.0 MiB/s         - c/B
   POLY1305 dec |      2.34 ns/B     408.1 MiB/s         - c/B
  POLY1305 auth |     0.238 ns/B      4003 MiB/s         - c/B
May 28 2022, 6:49 PM · patch, ppc, Feature Request, libgcrypt

May 27 2022

jukivili added a comment to T6006: Optimize Chacha20 and Poly1305 for PPC P10 LE.

-O2 problem with bench-slope seems strange. Does problem appear after this patch is applied?

May 27 2022, 7:15 PM · patch, ppc, Feature Request, libgcrypt

May 26 2022

werner triaged T6006: Optimize Chacha20 and Poly1305 for PPC P10 LE as Normal priority.
May 26 2022, 11:44 AM · patch, ppc, Feature Request, libgcrypt

May 17 2022

LFrank2021 added a comment to T5980: compilation error libgcrypt 1.10.1.

I do not claim I understand anything of this assembler syntax :)

May 17 2022, 11:14 AM · Testing, backport, ppc, AIX, libgcrypt, Bug Report
gniibe added a comment to T5980: compilation error libgcrypt 1.10.1.

For the second, I wonder if newer xlclang++ compiler works with 1.9.

May 17 2022, 4:58 AM · Testing, backport, ppc, AIX, libgcrypt, Bug Report
gniibe triaged T5980: compilation error libgcrypt 1.10.1 as Normal priority.

Thank you for the bug report.

May 17 2022, 4:31 AM · Testing, backport, ppc, AIX, libgcrypt, Bug Report

May 16 2022

werner added projects to T5980: compilation error libgcrypt 1.10.1: libgcrypt, AIX, ppc.
May 16 2022, 9:59 PM · Testing, backport, ppc, AIX, libgcrypt, Bug Report

Apr 19 2022

jukivili closed T5913: libgcrypt: bug fix for PPC bulk AES-GCM acceleratieration, missing HWF_PPC_ARCH_3_10 in HW feature as Resolved.
Apr 19 2022, 5:59 PM · ppc, libgcrypt

Apr 1 2022

dannytsen added a comment to T5913: libgcrypt: bug fix for PPC bulk AES-GCM acceleratieration, missing HWF_PPC_ARCH_3_10 in HW feature.

Hi Jussi, yes for some reason, it went missing, I was checking performance numbers and found out the line went missing. Thanks.

Apr 1 2022, 2:08 PM · ppc, libgcrypt
jukivili added a comment to T5913: libgcrypt: bug fix for PPC bulk AES-GCM acceleratieration, missing HWF_PPC_ARCH_3_10 in HW feature.

Fixed in master. I rechecked that bulk implementation passes tests with qemu-ppc64le.

Apr 1 2022, 8:55 AM · ppc, libgcrypt
jukivili added a comment to T5913: libgcrypt: bug fix for PPC bulk AES-GCM acceleratieration, missing HWF_PPC_ARCH_3_10 in HW feature.

Looks like that line went missing in third/final version of AES-GCM patch at https://dev.gnupg.org/T5700

Apr 1 2022, 8:51 AM · ppc, libgcrypt

Mar 31 2022

werner assigned T5913: libgcrypt: bug fix for PPC bulk AES-GCM acceleratieration, missing HWF_PPC_ARCH_3_10 in HW feature to jukivili.
Mar 31 2022, 10:46 PM · ppc, libgcrypt

Mar 2 2022

jukivili closed T5700: libgcrypt: bulk AES-GCM acceleration for ppc64le as Resolved.
Mar 2 2022, 8:24 PM · patch, ppc, libgcrypt, Feature Request

Dec 21 2021

jukivili added a comment to T5700: libgcrypt: bulk AES-GCM acceleration for ppc64le.

Ok, I'll add.

Dec 21 2021, 4:03 PM · patch, ppc, libgcrypt, Feature Request
werner added a comment to T5700: libgcrypt: bulk AES-GCM acceleration for ppc64le.

Seen. @jukivili can you please add it to the AUTHORS file?

Dec 21 2021, 10:13 AM · patch, ppc, libgcrypt, Feature Request

Dec 14 2021

dannytsen added a comment to T5700: libgcrypt: bulk AES-GCM acceleration for ppc64le.

Ok, I have subscribed to the mailing list. I have resent the DCO.

Dec 14 2021, 9:09 PM · patch, ppc, libgcrypt, Feature Request
jukivili added a comment to T5700: libgcrypt: bulk AES-GCM acceleration for ppc64le.

DCO has not appeared on mailing-list. You can this from check list archives, https://lists.gnupg.org/pipermail/gcrypt-devel/2021-December/thread.html

Dec 14 2021, 8:58 PM · patch, ppc, libgcrypt, Feature Request
dannytsen added a comment to T5700: libgcrypt: bulk AES-GCM acceleration for ppc64le.

Thanks Jussi, I did not receive the list moderator's email so I am not sure if the it has been posted on gcrypt-devel@gnupg.org. If not, I can resend the DCO. Thanks.

Dec 14 2021, 7:58 PM · patch, ppc, libgcrypt, Feature Request
jukivili added a comment to T5700: libgcrypt: bulk AES-GCM acceleration for ppc64le.

I did some finishing touches on coding style:

Dec 14 2021, 7:39 PM · patch, ppc, libgcrypt, Feature Request

Dec 13 2021

dannytsen added a comment to T5700: libgcrypt: bulk AES-GCM acceleration for ppc64le.

Hi Jussi,

Dec 13 2021, 7:54 PM · patch, ppc, libgcrypt, Feature Request

Dec 12 2021

jukivili added a comment to T5700: libgcrypt: bulk AES-GCM acceleration for ppc64le.

Few comments on new patch:

Dec 12 2021, 7:18 PM · patch, ppc, libgcrypt, Feature Request

Dec 10 2021

dannytsen added a comment to T5700: libgcrypt: bulk AES-GCM acceleration for ppc64le.

Hi jukivili,

Dec 10 2021, 3:21 PM · patch, ppc, libgcrypt, Feature Request

Dec 7 2021

dannytsen added a comment to T5700: libgcrypt: bulk AES-GCM acceleration for ppc64le.

Hi jukivili,
I ran some basic tests and it did show the errors. I am in the process investigating what went wrong. In the meantime, i also included test result that I have used in my testing from bench-slope. In this test, I captured the message with 272 bytes buffer from the original libgcrypt repo and my optimized repo. Note that the bulk version of my code do 8x unrolling and the rest will do 16 bytes. So the first 2 128 bytes ran thru gcry_ppc_aes_gcm_encrypt and the rest of the 16 bytes thru gcm_ctr_encrypt (cipher-gcm.c).

Dec 7 2021, 5:36 PM · patch, ppc, libgcrypt, Feature Request

Dec 6 2021

dannytsen added a comment to T5700: libgcrypt: bulk AES-GCM acceleration for ppc64le.

Thanks jukivili for the review.

Dec 6 2021, 3:37 PM · patch, ppc, libgcrypt, Feature Request

Dec 4 2021

jukivili added a comment to T5700: libgcrypt: bulk AES-GCM acceleration for ppc64le.

Thanks, however I didn't see your email on mailing-list. Maybe the email got stuck on the way.

Dec 4 2021, 12:46 PM · patch, ppc, libgcrypt, Feature Request

Dec 2 2021

dannytsen added a comment to T5700: libgcrypt: bulk AES-GCM acceleration for ppc64le.

I sent a copy to gcrypt-devel@gnupg.org. Hope this is the right process. Thanks.

Dec 2 2021, 5:47 PM · patch, ppc, libgcrypt, Feature Request
jukivili added a comment to T5700: libgcrypt: bulk AES-GCM acceleration for ppc64le.

Please read doc/HACKING carefully on the process of sending DCO the right way.

Dec 2 2021, 5:31 PM · patch, ppc, libgcrypt, Feature Request

Nov 23 2021

dannytsen added a comment to T5700: libgcrypt: bulk AES-GCM acceleration for ppc64le.

Hi Werner, Here is the DCO. Thanks.

Nov 23 2021, 3:51 PM · patch, ppc, libgcrypt, Feature Request
werner triaged T5700: libgcrypt: bulk AES-GCM acceleration for ppc64le as Normal priority.

FWIW: We need a DCO; see doc/HACKING.

Nov 23 2021, 9:06 AM · patch, ppc, libgcrypt, Feature Request

Apr 1 2021

werner created ppc.
Apr 1 2021, 11:06 AM