Add builtins for f16/f128 float conversions
#593
There are no files selected for viewing
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
| Original file line number | Diff line number | Diff line change |
|---|---|---|
|
|
@@ -52,8 +52,10 @@ where | |
| // destination format. We can convert by simply right-shifting with | ||
| // rounding and adjusting the exponent. | ||
| abs_result = (a_abs >> sign_bits_delta).cast(); | ||
| // Cast before shifting to prevent overflow. | ||
|
There was a problem hiding this comment. I'm having trouble understand how exactly this helps prevent overflows. Aren't you casting to a smaller size (u16) here instead of operating as a u32? There was a problem hiding this comment. @Amanieu This is indeed not an issue for truncation to |
||
| let bias_diff: R::Int = src_exp_bias.wrapping_sub(dst_exp_bias).cast(); | ||
| let tmp = bias_diff << R::SIGNIFICAND_BITS; | ||
| abs_result = abs_result.wrapping_sub(tmp); | ||
|
|
||
| let round_bits = a_abs & round_mask; | ||
| if round_bits > halfway { | ||
|
|
@@ -67,13 +69,17 @@ where | |
| // a is NaN. | ||
| // Conjure the result by beginning with infinity, setting the qNaN | ||
| // bit and inserting the (truncated) trailing NaN field. | ||
| // Cast before shifting to prevent overflow. | ||
| let dst_inf_exp: R::Int = dst_inf_exp.cast(); | ||
| abs_result = dst_inf_exp << R::SIGNIFICAND_BITS; | ||
| abs_result |= dst_qnan; | ||
| abs_result |= dst_nan_code | ||
| & ((a_abs & src_nan_code) >> (F::SIGNIFICAND_BITS - R::SIGNIFICAND_BITS)).cast(); | ||
| } else if a_abs >= overflow { | ||
| // a overflows to infinity. | ||
| // Cast before shifting to prevent overflow. | ||
| let dst_inf_exp: R::Int = dst_inf_exp.cast(); | ||
| abs_result = dst_inf_exp << R::SIGNIFICAND_BITS; | ||
| } else { | ||
| // a underflows on conversion to the destination type or is an exact | ||
| // zero. The result may be a denormal or zero. Extract the exponent | ||
|
|
@@ -124,3 +130,44 @@ intrinsics! { | |
| a as f32 | ||
| } | ||
| } | ||
|
|
||
| #[cfg(not(feature = "no-f16-f128"))] | ||
| intrinsics! { | ||
| #[avr_skip] | ||
| #[aapcs_on_arm] | ||
| #[arm_aeabi_alias = __aeabi_f2h] | ||
| pub extern "C" fn __truncsfhf2(a: f32) -> f16 { | ||
| trunc(a) | ||
| } | ||
|
|
||
| #[avr_skip] | ||
| #[aapcs_on_arm] | ||
| pub extern "C" fn __gnu_f2h_ieee(a: f32) -> f16 { | ||
| trunc(a) | ||
| } | ||
|
|
||
| #[avr_skip] | ||
| #[aapcs_on_arm] | ||
| #[arm_aeabi_alias = __aeabi_d2h] | ||
| pub extern "C" fn __truncdfhf2(a: f64) -> f16 { | ||
| trunc(a) | ||
| } | ||
|
|
||
| #[avr_skip] | ||
| #[aapcs_on_arm] | ||
| pub extern "C" fn __trunctfhf2(a: f128) -> f16 { | ||
| trunc(a) | ||
| } | ||
|
|
||
| #[avr_skip] | ||
| #[aapcs_on_arm] | ||
| pub extern "C" fn __trunctfsf2(a: f128) -> f32 { | ||
| trunc(a) | ||
| } | ||
|
|
||
| #[avr_skip] | ||
| #[aapcs_on_arm] | ||
| pub extern "C" fn __trunctfdf2(a: f128) -> f64 { | ||
| trunc(a) | ||
| } | ||
| } | ||
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Oops, something went wrong.
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Features should usually be additive and not take things away -- so, can't we have an "f16-f128" feature that enables these intrinsics, and have them off by default?
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
I think that was just following the
no-asmfeature above, but they could probably both be flipped to be additiveThere was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
This "adds" support for codegen backends that don't support f16/f128. If a
f16-f128feature is used instead, liballoc would likely enable it when depending on compiler-builtins, which makes it impossible to disable the f16/f128 intrinsics using--featureswhen invoking cargo, instead requiring the standard library sources to be patched.There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
... what?^^ I am confused. "Additive" is always meant in a semver way when it comes to cargo feature: if a dependent builds without the feature, it should also build with the feature. Don't try to re-define well-established terms, please. The leading "no" is also a dead giveaway that this not a negative feature, removing things, rather than adding them.
We are currently (once that PR lands) setting this feature in
stdbecause of said codegen backends. But that means buildingcorealone will still fail with these codegen backends. It all makes little sense to me.That sounds wrong to me. Instead, bootstrap should enable the feature when it deems them supported by the environment.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
The build system of these codegen backends should set it, not libstd. For example for no-asm, this is what cg_clif does when building the standard library:
https://github.com/rust-lang/rustc_codegen_cranelift/blob/6db27529ddbb6d94e0c990e527aa8341c3d85a79/build_system/build_sysroot.rs#L279
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
I can accept the answer "this is a hack, compiler_builtins is special and has a single consumer and so we can get away with negative features".