Skip to content

icache: support early refill forwarding and lookups on refill#714

Draft
awariac wants to merge 6 commits intokuznia-rdzeni:masterfrom
awariac:piotro/icache-refill-lookup
Draft

icache: support early refill forwarding and lookups on refill#714
awariac wants to merge 6 commits intokuznia-rdzeni:masterfrom
awariac:piotro/icache-refill-lookup

Conversation

@awariac
Copy link
Copy Markdown
Member

@awariac awariac commented Jun 25, 2024

Closes #286

awariac added 3 commits June 25, 2024 22:56
- result would be available in the same cycle as on bus.
@awariac awariac added performance Improves performance benchmark Benchmarks should be run for this change and removed benchmark Benchmarks should be run for this change labels Jun 25, 2024
@github-actions
Copy link
Copy Markdown

Benchmarks summary

Performance benchmarks

aha-mont64 crc32 minver nettle-sha256 nsichneu slre statemate ud
0.418 0.513 0.338 0.656 0.369 0.29 0.327 0.433

You can view all the metrics here.

Synthesis benchmarks (basic)

Device utilisation: (ECP5) LUTs used as DFF: (ECP5) LUTs used as carry: (ECP5) LUTs used as ram: (ECP5) Max clock frequency (Fmax)
21850 6009 802 976 49

Synthesis benchmarks (full)

Device utilisation: (ECP5) LUTs used as DFF: (ECP5) LUTs used as carry: (ECP5) LUTs used as ram: (ECP5) Max clock frequency (Fmax)
37472 9421 1944 1156 39

@kuznia-rdzeni kuznia-rdzeni deleted a comment from github-actions Bot Jun 25, 2024
@awariac awariac force-pushed the piotro/icache-refill-lookup branch from db1ad43 to cdbd1b0 Compare June 25, 2024 22:57
@github-actions
Copy link
Copy Markdown

Benchmarks summary

Performance benchmarks

aha-mont64 crc32 minver nettle-sha256 nsichneu slre statemate ud
0.418 0.513 0.338 0.656 0.371 0.29 0.327 0.433

You can view all the metrics here.

Synthesis benchmarks (basic)

Device utilisation: (ECP5) LUTs used as DFF: (ECP5) LUTs used as carry: (ECP5) LUTs used as ram: (ECP5) Max clock frequency (Fmax)
25675 6172 770 976 51

Synthesis benchmarks (full)

Device utilisation: (ECP5) LUTs used as DFF: (ECP5) LUTs used as carry: (ECP5) LUTs used as ram: (ECP5) Max clock frequency (Fmax)
29543 9417 1976 1152 35

@github-actions
Copy link
Copy Markdown

Benchmarks summary

Performance benchmarks

aha-mont64 crc32 minver nettle-sha256 nsichneu slre statemate ud
0.412 (0.000) 0.539 (0.000) ▲ 0.358 (+0.001) ▲ 0.642 (+0.000) ▲ 0.360 (+0.009) ▲ 0.292 (+0.000) ▲ 0.328 (+0.000) 0.434 (0.000)

You can view all the metrics here.

Synthesis benchmarks (basic)

Device utilisation: (ECP5) LUTs used as DFF: (ECP5) LUTs used as carry: (ECP5) LUTs used as ram: (ECP5) Max clock frequency (Fmax)
▲ 21504 (+1049) ▲ 6976 (+245) ▲ 1416 (+32) ▲ 1540 (+4) ▼ 48 (-7)

Synthesis benchmarks (full)

Device utilisation: (ECP5) LUTs used as DFF: (ECP5) LUTs used as carry: (ECP5) LUTs used as ram: (ECP5) Max clock frequency (Fmax)
▼ 34066 (-1448) ▲ 9469 (+239) ▲ 1900 (+32) 2064 (0) ▼ 39 (-3)

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

benchmark Benchmarks should be run for this change performance Improves performance

Projects

None yet

Development

Successfully merging this pull request may close these issues.

Instruction cache doesn't do lookups while refilling

1 participant