search

riscvarchive / riscv-v-spec

Working draft of the proposed RISC-V V vector extension
https://jira.riscv.org/browse/RVG-122
Creative Commons Attribution 4.0 International
973 stars 272 forks source link

The behavior of slideup/slidedown when i is 0 #940

Closed SoufSilence closed 10 months ago

SoufSilence commented 10 months ago

The spec defines the behavior of slideup/slidedown when i>0, but it does not seem to define the behavior when i=0. Is this an oversight?

For slideup: 图片 For slidedown: 图片

aswaterman commented 10 months ago

Normally, vstart is 0. The case below the one you highlighted addresses the 0 case. The case you highlighted is relevant only when vstart>0.

On Thu, Jan 25, 2024 at 7:29 PM SoufSilence @.***> wrote:

The spec defines the behavior of slideup/slidedown when i>0, but it does not seem to define the behavior when i=0. Is this an oversight?

For slideup: default.png (view on web) https://github.com/riscv/riscv-v-spec/assets/13818622/c774bb1d-5ef6-42a8-af9f-750aad56ec98 For slidedown: default.png (view on web) https://github.com/riscv/riscv-v-spec/assets/13818622/e15c43e3-8ff3-4814-904e-e3e307450a12

— Reply to this email directly, view it on GitHub https://github.com/riscv/riscv-v-spec/issues/940, or unsubscribe https://github.com/notifications/unsubscribe-auth/AAH3XQTYFEAVJVE2BX7WKRLYQMPI5AVCNFSM6AAAAABCLRJEPCVHI2DSMVQWIX3LMV43ASLTON2WKOZSGEYDCNJTGA4TIOI . You are receiving this because you are subscribed to this thread.Message ID: @.***>

SoufSilence commented 10 months ago

Normally, vstart is 0. The case below the one you highlighted addresses the 0 case. The case you highlighted is relevant only when vstart>0. … On Thu, Jan 25, 2024 at 7:29 PM SoufSilence @.> wrote: The spec defines the behavior of slideup/slidedown when i>0, but it does not seem to define the behavior when i=0. Is this an oversight? For slideup: default.png (view on web) https://github.com/riscv/riscv-v-spec/assets/13818622/c774bb1d-5ef6-42a8-af9f-750aad56ec98 For slidedown: default.png (view on web) https://github.com/riscv/riscv-v-spec/assets/13818622/e15c43e3-8ff3-4814-904e-e3e307450a12 — Reply to this email directly, view it on GitHub <#940>, or unsubscribe https://github.com/notifications/unsubscribe-auth/AAH3XQTYFEAVJVE2BX7WKRLYQMPI5AVCNFSM6AAAAABCLRJEPCVHI2DSMVQWIX3LMV43ASLTON2WKOZSGEYDCNJTGA4TIOI . You are receiving this because you are subscribed to this thread.Message ID: @.>

Thank you very much for your quick reply! But please forgive me for not understanding your comment. Maybe I didn't express myself clearly. What I want to say is that when vstart is 0 and OFFSET is 5, for example, the spec only specifies that vd[1], vd[2], vd[3] and vd[4] should remain unchanged, but the value of vd[0] is not specified.

aswaterman commented 10 months ago

Ah, it looks like the issue is that we are looking at different versions of the spec. If you look at the HEAD of master, you'll see that the < was changed to <= a few years ago.

SoufSilence commented 10 months ago

Ah, it looks like the issue is that we are looking at different versions of the spec. If you look at the HEAD of master, you'll see that the < was changed to <= a few years ago.

Yes we looked at different versions, now I know the way to check the latest version, thanks for your reply.