This is an archive of the discontinued LLVM Phabricator instance.

Document rounding for llvm.lround and llvm.lrint
ClosedPublic

Authored by kpn on Oct 10 2019, 9:34 AM.

Download Raw Diff

Details

Reviewers

hfinkel
andrew.w.kaylor
craig.topper
cameron.mcinally
efriedma

Commits

rGd2b6cc7ff6d8: Document more specifically the rounding for "llvm.round".

Summary

The handling of rounding is not fully documented for the llvm.lround and llvm.lrint family of intrinsics.

This documentation was requested during the reviews of D64746.

Diff Detail

Repository: rG LLVM Github Monorepo

Event Timeline

kpn created this revision.Oct 10 2019, 9:34 AM

Herald added a project: Restricted Project. · View Herald TranscriptOct 10 2019, 9:34 AM

Herald added a subscriber: llvm-commits. · View Herald Transcript

kpn added reviewers: andrew.w.kaylor, craig.topper, cameron.mcinally, efriedma.Nov 11 2019, 10:11 AM

These intrinsics don't raise exceptions, semantically; they're readnone.

Drop the verbiage about exceptions. Update ticket description and title.

craig.topper added inline comments.Nov 13 2019, 12:19 PM

llvm/docs/LangRef.rst
12788–12790	Doesn't lrint use the current rounding mode?

efriedma added inline comments.Nov 13 2019, 12:30 PM

llvm/docs/LangRef.rst
12788–12790	Sort of? I mean, in this context, we assume the "current rounding mode" is round-to-nearest-even. Probably worth clarifying what we mean by "nearest" here.

kpn marked an inline comment as done.Nov 13 2019, 12:40 PM

kpn added inline comments.

llvm/docs/LangRef.rst
12788–12790	We specify how rounding is done very rarely, and when we do we usually say "nearest integer" without mentioning the rounding mode. Do we want to change all references to "nearest integer" to state that the current rounding mode is used? Do we want a mention somewhere else to cover all uses?

efriedma added inline comments.Nov 13 2019, 12:55 PM

llvm/docs/LangRef.rst
12788–12790	We only use "nearest integer" in connection to rint/round/etc. In that context, it's not obvious to a casual reader what this means in the context of IEEE arithmetic. In other contexts, floating-point arithmetic uses the default floating-point rounding mode; this is stated explicitly at http://llvm.org/docs/LangRef.html#floating-point-environment .

kpn marked an inline comment as done.Nov 13 2019, 1:00 PM

kpn added inline comments.

llvm/docs/LangRef.rst
12788–12790	Looking more closely, the lrint parts of this patch are just grammar changes. I could commit those tomorrow and this patch would be just lround. The *lround part of the patch is explicit in stating the rounding. Does that seem reasonable?

efriedma accepted this revision.Nov 13 2019, 2:48 PM

efriedma added inline comments.

llvm/docs/LangRef.rst
12788–12790	Fine. In that case, LGTM

This revision is now accepted and ready to land.Nov 13 2019, 2:48 PM

Closed by commit rGd2b6cc7ff6d8: Document more specifically the rounding for "llvm.round". (authored by kpn). · Explain WhyNov 14 2019, 10:24 AM

This revision was automatically updated to reflect the committed changes.

Revision Contents

Path

Size

llvm/

docs/

LangRef.rst

4 lines

Diff 229347

llvm/docs/LangRef.rst

This file is larger than 256 KB, so syntax highlighting is disabled by default.

Show First 20 Lines • Show All 12,703 Lines • ▼ Show 20 Lines	::
declare i64 @llvm.lround.i64.f80(float %Val)		declare i64 @llvm.lround.i64.f80(float %Val)
declare i64 @llvm.lround.i64.f128(double %Val)		declare i64 @llvm.lround.i64.f128(double %Val)
declare i64 @llvm.lround.i64.ppcf128(double %Val)		declare i64 @llvm.lround.i64.ppcf128(double %Val)

Overview:		Overview:
"""""""""		"""""""""

The '``llvm.lround.*``' intrinsics return the operand rounded to the nearest		The '``llvm.lround.*``' intrinsics return the operand rounded to the nearest
integer.		integer with ties away from zero.


Arguments:		Arguments:
""""""""""		""""""""""

The argument is a floating-point number and the return value is an integer		The argument is a floating-point number and the return value is an integer
type.		type.

Show All 19 Lines	::
declare i64 @llvm.lround.i64.f80(float %Val)		declare i64 @llvm.lround.i64.f80(float %Val)
declare i64 @llvm.lround.i64.f128(double %Val)		declare i64 @llvm.lround.i64.f128(double %Val)
declare i64 @llvm.lround.i64.ppcf128(double %Val)		declare i64 @llvm.lround.i64.ppcf128(double %Val)

Overview:		Overview:
"""""""""		"""""""""

The '``llvm.llround.*``' intrinsics return the operand rounded to the nearest		The '``llvm.llround.*``' intrinsics return the operand rounded to the nearest
integer.		integer with ties away from zero.

Arguments:		Arguments:
""""""""""		""""""""""

The argument is a floating-point number and the return value is an integer		The argument is a floating-point number and the return value is an integer
type.		type.

Semantics:		Semantics:
Show All 23 Lines	::
declare i64 @llvm.lrint.i64.f64(double %Val)		declare i64 @llvm.lrint.i64.f64(double %Val)
declare i64 @llvm.lrint.i64.f80(float %Val)		declare i64 @llvm.lrint.i64.f80(float %Val)
declare i64 @llvm.lrint.i64.f128(double %Val)		declare i64 @llvm.lrint.i64.f128(double %Val)
declare i64 @llvm.lrint.i64.ppcf128(double %Val)		declare i64 @llvm.lrint.i64.ppcf128(double %Val)

Overview:		Overview:
"""""""""		"""""""""

The '``llvm.lrint.*``' intrinsics return the operand rounded to the nearest		The '``llvm.lrint.*``' intrinsics return the operand rounded to the nearest
integer.		integer.

		craig.topperUnsubmitted Not Done Reply Inline Actions Doesn't lrint use the current rounding mode? craig.topper: Doesn't lrint use the current rounding mode?
		efriedmaUnsubmitted Not Done Reply Inline Actions Sort of? I mean, in this context, we assume the "current rounding mode" is round-to-nearest-even. Probably worth clarifying what we mean by "nearest" here. efriedma: Sort of? I mean, in this context, we assume the "current rounding mode" is round-to-nearest…
		kpnAuthorUnsubmitted Done Reply Inline Actions We specify how rounding is done very rarely, and when we do we usually say "nearest integer" without mentioning the rounding mode. Do we want to change all references to "nearest integer" to state that the current rounding mode is used? Do we want a mention somewhere else to cover all uses? kpn: We specify how rounding is done very rarely, and when we do we usually say "nearest integer"…
		efriedmaUnsubmitted Not Done Reply Inline Actions We only use "nearest integer" in connection to rint/round/etc. In that context, it's not obvious to a casual reader what this means in the context of IEEE arithmetic. In other contexts, floating-point arithmetic uses the default floating-point rounding mode; this is stated explicitly at http://llvm.org/docs/LangRef.html#floating-point-environment . efriedma: We only use "nearest integer" in connection to rint/round/etc. In that context, it's not…
		kpnAuthorUnsubmitted Done Reply Inline Actions Looking more closely, the lrint parts of this patch are just grammar changes. I could commit those tomorrow and this patch would be just lround. The lround part of the patch is explicit in stating the rounding. Does that seem reasonable? kpn:* Looking more closely, the *lrint parts of this patch are just grammar changes. I could commit…
		efriedmaUnsubmitted Not Done Reply Inline Actions Fine. In that case, LGTM efriedma: Fine. In that case, LGTM

Arguments:		Arguments:
""""""""""		""""""""""

The argument is a floating-point number and the return value is an integer		The argument is a floating-point number and the return value is an integer
type.		type.

Semantics:		Semantics:
▲ Show 20 Lines • Show All 5,205 Lines • Show Last 20 Lines