Stream: git-wasmtime

Topic: wasmtime / issue #3523 Pathological time spent in registe...

Wasmtime GitHub notifications bot (Nov 15 2021 at 21:11):

alexcrichton edited issue #3523:

We have implemented this repo https://github.com/girazoki/wasmtime-try to try something that has been happening us with some latest code change we have done at moonbeam. The code change is https://github.com/PureStake/moonbeam/pull/989.

Without the PR, wasmtime takes around 15 seconds to compile the wasm. However, if we include the change (removal of the trait) wasmtime takes less than a second to compile. We wonder how these differences can be so big with such a small code change.

Both runtimes are in https://github.com/girazoki/wasmtime-try to try them out
Any help would be appreciated, thanks!

Wasmtime GitHub notifications bot (Nov 15 2021 at 21:12):

cfallin commented on issue #3523:

@alexcrichton That's a good question; ordinarily we'd definitely want to chase something like this, but if the RA2 transition happens soon-ish then the payoff of the time it would take to understand the issue and update the old regalloc, and be confident in the fix, doesn't seem worth it. Especially as many of the details of regalloc.rs have fallen out of my mental L2 cache by now...

These sorts of things do put pressure on the transition, though, so hopefully the moving pieces needed for that will come together soon!

Wasmtime GitHub notifications bot (Nov 15 2021 at 21:16):

alexcrichton commented on issue #3523:

@adamrk given your interest on this issue you might be interested in this one as well (detailed here) but if not no worries!

Wasmtime GitHub notifications bot (Nov 16 2021 at 14:11):

crystalin commented on issue #3523:

Not sure if that would help, but here are the perf (you can open in https://www.speedscope.app/ ) for when it is loading the wasm: https://drive.google.com/file/d/1LSid_h9vs8LGtbaQ8z41CD3vn4-y4g2U/view?usp=sharing

(Here is where most of the work goes it seems)
.

Wasmtime GitHub notifications bot (Nov 16 2021 at 14:11):

crystalin edited a comment on issue #3523:

Not sure if that would help, but here are the perf (you can open in https://www.speedscope.app/ ) for when it is loading the wasm: https://drive.google.com/file/d/1LSid_h9vs8LGtbaQ8z41CD3vn4-y4g2U/view?usp=sharing

(Here is where most of the work goes it seems)

Wasmtime GitHub notifications bot (Nov 16 2021 at 14:11):

crystalin edited a comment on issue #3523:

Not sure if that would help, but here are the perf (you can open in https://www.speedscope.app/ ) for when it is loading the wasm: https://drive.google.com/file/d/1LSid_h9vs8LGtbaQ8z41CD3vn4-y4g2U/view?usp=sharing

(Here is where most of the work goes it seems
)

Wasmtime GitHub notifications bot (Nov 16 2021 at 14:11):

crystalin edited a comment on issue #3523:

Not sure if that would help, but here are the perf (you can open in https://www.speedscope.app/ ) for when it is loading the wasm: https://drive.google.com/file/d/1LSid_h9vs8LGtbaQ8z41CD3vn4-y4g2U/view?usp=sharing

(Here is where most of the work goes it seems)

Wasmtime GitHub notifications bot (Nov 19 2021 at 16:22):

alexcrichton labeled issue #3523:

We have implemented this repo https://github.com/girazoki/wasmtime-try to try something that has been happening us with some latest code change we have done at moonbeam. The code change is https://github.com/PureStake/moonbeam/pull/989.

Without the PR, wasmtime takes around 15 seconds to compile the wasm. However, if we include the change (removal of the trait) wasmtime takes less than a second to compile. We wonder how these differences can be so big with such a small code change.

Both runtimes are in https://github.com/girazoki/wasmtime-try to try them out
Any help would be appreciated, thanks!

Wasmtime GitHub notifications bot (Nov 19 2021 at 16:23):

alexcrichton labeled issue #3523:

We have implemented this repo https://github.com/girazoki/wasmtime-try to try something that has been happening us with some latest code change we have done at moonbeam. The code change is https://github.com/PureStake/moonbeam/pull/989.

Without the PR, wasmtime takes around 15 seconds to compile the wasm. However, if we include the change (removal of the trait) wasmtime takes less than a second to compile. We wonder how these differences can be so big with such a small code change.

Both runtimes are in https://github.com/girazoki/wasmtime-try to try them out
Any help would be appreciated, thanks!

Wasmtime GitHub notifications bot (Nov 26 2021 at 22:57):

adamrk commented on issue #3523:

Sorry, just noticed this. Yeah, I'll take a look at it.

Wasmtime GitHub notifications bot (Dec 15 2021 at 11:32):

JoshOrndorff commented on issue #3523:

@alexcrichton We have hit what appears to be the same issue in the Moonbeam project again. This time it is in this wasm file.

Following your example, @girazoki, explored this:

RUST_LOG=wasmtime_cranelift wasmtime compile moonbase_runtime.wasm

...

 DEBUG wasmtime_cranelift::compiler > FuncIndex(1276) translated in 13.571936365s
 TRACE wasmtime_cranelift::compiler > FuncIndex(1276) timing info
======== ========  ==================================
   Total     Self  Pass
-------- --------  ----------------------------------
   0.510    0.510  Translate WASM function
  13.056    0.030  Compilation passes
   0.001    0.001  Control flow graph
   0.002    0.002  Dominator tree
   0.000    0.000  Loop analysis
   0.001    0.001  Pre-legalization rewriting
   0.005    0.005  Dead code elimination
   0.009    0.009  Global value numbering
   0.044    0.033  Loop invariant code motion
   0.000    0.000  Remove unreachable blocks
   0.210    0.210  Remove constant phi-nodes
   0.222    0.222  VCode lowering
   0.060    0.060  VCode post-register allocation finalization
   0.029    0.029  VCode emission
   0.000    0.000  VCode emission finalization
  12.443   12.443  Register allocation
   0.006    0.006  Binary machine code emission
   0.009    0.009  Layout full renumbering
======== ========  ==================================

How did you figure out the human-readale or even the mangled name of the function to know where to put the #[inline(never)]?

Wasmtime GitHub notifications bot (Dec 15 2021 at 17:49):

alexcrichton commented on issue #3523:

Oh nice, thanks for that! I was able to find the function-in-question with:

RUST_LOG=wasmtime_cranelift=debug ./target/release/wasmtime compile moonbase_runtime.compact.wasm

which prints out:

...
 DEBUG wasmtime_cranelift::compiler > FuncIndex(1776) translated in 70.436892ms
 DEBUG wasmtime_cranelift::compiler > FuncIndex(679) translated in 69.31585ms
 DEBUG wasmtime_cranelift::compiler > FuncIndex(2050) translated in 61.266995ms
 DEBUG wasmtime_cranelift::compiler > FuncIndex(1574) translated in 68.276648ms
 DEBUG wasmtime_cranelift::compiler > FuncIndex(1622) translated in 75.956103ms
 DEBUG wasmtime_cranelift::compiler > FuncIndex(1983) translated in 85.31272ms
 DEBUG wasmtime_cranelift::compiler > FuncIndex(1347) translated in 282.235891ms
 DEBUG wasmtime_cranelift::compiler > FuncIndex(1276) translated in 19.01582367s

which points to function 1276 as the offender here.

Using wasm-opt I got:

$ wasm-opt --extract-function-index=1276 ./moonbase_runtime.compact.wasm -o extract.wasm
extracting <xcm_executor::XcmExecutor<Config>\20as\20xcm::v2::traits::ExecuteXcm<<Config\20as\20xcm_executor::config::Config>::Call>>::execute_xcm_in_credit::hf77207caf74f0f13

which I can confirm the extracted wasm file takes ~20s to compile, quite too long!

Wasmtime GitHub notifications bot (Dec 15 2021 at 17:50):

alexcrichton edited issue #3523:

Original Issue

We have implemented this repo https://github.com/girazoki/wasmtime-try to try something that has been happening us with some latest code change we have done at moonbeam. The code change is https://github.com/PureStake/moonbeam/pull/989.

Without the PR, wasmtime takes around 15 seconds to compile the wasm. However, if we include the change (removal of the trait) wasmtime takes less than a second to compile. We wonder how these differences can be so big with such a small code change.

Both runtimes are in https://github.com/girazoki/wasmtime-try to try them out
Any help would be appreciated, thanks!

Current status

Modules with way-too-long-compile-times:

• https://github.com/bytecodealliance/wasmtime/issues/3523#issuecomment-969324846
• https://github.com/bytecodealliance/wasmtime/issues/3523#issuecomment-995023900

Wasmtime GitHub notifications bot (Apr 26 2022 at 15:01):

alexcrichton commented on issue #3523:

I'm going to close this given the discussion on https://github.com/bytecodealliance/wasmtime/issues/4060. The new regalloc probably improves things here but these sorts of outliers are inevitable with Cranelift's design goals.

Wasmtime GitHub notifications bot (Apr 26 2022 at 15:01):

alexcrichton closed issue #3523:

Original Issue

We have implemented this repo https://github.com/girazoki/wasmtime-try to try something that has been happening us with some latest code change we have done at moonbeam. The code change is https://github.com/PureStake/moonbeam/pull/989.

Without the PR, wasmtime takes around 15 seconds to compile the wasm. However, if we include the change (removal of the trait) wasmtime takes less than a second to compile. We wonder how these differences can be so big with such a small code change.

Both runtimes are in https://github.com/girazoki/wasmtime-try to try them out
Any help would be appreciated, thanks!

Current status

Modules with way-too-long-compile-times:

• https://github.com/bytecodealliance/wasmtime/issues/3523#issuecomment-969324846
• https://github.com/bytecodealliance/wasmtime/issues/3523#issuecomment-995023900

Last updated: Oct 23 2024 at 20:03 UTC