Stream: git-cranelift

Topic: cranelift / PR #1391 Add maximum threshold for number of ...

GitHub (Feb 13 2020 at 19:28):

nalmt opened PR #1391 from pathological_case_951 to master:

To fix this case that may take forever to compile (issue #951):
function %a() {
ebb477777777:
}

We decide to define a maximum threshold for the number of ebbs in functions.

Based on a large WASM program (https://github.com/mozilla/perf-automation/blob/master/benchmarks/wasm-misc/AngryBots.wasm), its IR functions does not exceed 1414 ebbs.
A number 100 times greater seems (currently) enough to define our maximum threshold.

static MAX_EBBS_IN_A_FUNCTION: u32 = 100000;

if ebb_num.as_u32() > MAX_EBBS_IN_A_FUNCTION {
return Err(self.error("too many ebbs"));
}

GitHub (Feb 13 2020 at 19:31):

nalmt requested bnjbvr for a review on PR #1391.

GitHub (Feb 14 2020 at 15:09):

nalmt updated PR #1391 from pathological_case_951 to master:

To fix this case that may take forever to compile (issue #951):
function %a() {
ebb477777777:
}

We decide to define a maximum threshold for the number of ebbs in functions.

Based on a large WASM program (https://github.com/mozilla/perf-automation/blob/master/benchmarks/wasm-misc/AngryBots.wasm), its IR functions does not exceed 1414 ebbs.
A number 100 times greater seems (currently) enough to define our maximum threshold.

static MAX_EBBS_IN_A_FUNCTION: u32 = 100000;

if ebb_num.as_u32() > MAX_EBBS_IN_A_FUNCTION {
return Err(self.error("too many ebbs"));
}

GitHub (Feb 14 2020 at 15:11):

nalmt updated PR #1391 from pathological_case_951 to master:

To fix this case that may take forever to compile (issue #951):
function %a() {
ebb477777777:
}

We decide to define a maximum threshold for the number of ebbs in functions.

Based on a large WASM program (https://github.com/mozilla/perf-automation/blob/master/benchmarks/wasm-misc/AngryBots.wasm), its IR functions does not exceed 1414 ebbs.
A number 100 times greater seems (currently) enough to define our maximum threshold.

static MAX_EBBS_IN_A_FUNCTION: u32 = 100000;

if ebb_num.as_u32() > MAX_EBBS_IN_A_FUNCTION {
return Err(self.error("too many ebbs"));
}

GitHub (Feb 14 2020 at 15:12):

nalmt updated PR #1391 from pathological_case_951 to master:

To fix this case that may take forever to compile (issue #951):
function %a() {
ebb477777777:
}

We decide to define a maximum threshold for the number of ebbs in functions.

Based on a large WASM program (https://github.com/mozilla/perf-automation/blob/master/benchmarks/wasm-misc/AngryBots.wasm), its IR functions does not exceed 1414 ebbs.
A number 100 times greater seems (currently) enough to define our maximum threshold.

static MAX_EBBS_IN_A_FUNCTION: u32 = 100000;

if ebb_num.as_u32() > MAX_EBBS_IN_A_FUNCTION {
return Err(self.error("too many ebbs"));
}

GitHub (Feb 14 2020 at 15:25):

nalmt updated PR #1391 from pathological_case_951 to master:

To fix this case that may take forever to compile (issue #951):
function %a() {
ebb477777777:
}

We decide to define a maximum threshold for the number of ebbs in functions.

Based on a large WASM program (https://github.com/mozilla/perf-automation/blob/master/benchmarks/wasm-misc/AngryBots.wasm), its IR functions does not exceed 1414 ebbs.
A number 100 times greater seems (currently) enough to define our maximum threshold.

static MAX_EBBS_IN_A_FUNCTION: u32 = 100000;

if ebb_num.as_u32() > MAX_EBBS_IN_A_FUNCTION {
return Err(self.error("too many ebbs"));
}

GitHub (Feb 14 2020 at 15:46):

nalmt closed without merge PR #1391.

Last updated: Dec 23 2024 at 13:07 UTC