[flang] Add -ffast-math and -Ofast

[tblah]

Summary

RFC: https://discourse.llvm.org/t/rfc-the-meaning-of-ofast/66554

-Ofast is defined as -O3 -ffast-math

-ffast-math will mean the following:

• -fno-honor-infinities
• -fno-honor-nans
• -fassociative-math
• -freciprocal-math
• -fapprox-func
• -fno-signed-zeros
• -ffp-contract=fast

This definition matches the meaning of the "fast" attribute in LLVM IR: https://llvm.org/docs/LangRef.html#fast-math-flags

-fstack-arrays should also be enabled with -Ofast but that flang is not yet implemented and so has been left for a future patch.

clang -cc1 accepts -Ofast. I did not add -Ofast to flang-new -fc1 because it is redundant because the compiler driver will always resolve -Ofast into -O3 -ffast-math (I added a test for this in flang/test/Driver/fast_math.f90; there is a test that -ffast-math leads to the fast attribute in flang/test/Lower/fast-math-arithmetic.f90). This means that flang-new -Ofast will be forwarded as flang-new -fc1 -O3 -ffast-math, which is entirely equivalent. Avoiding adding the umbrella flag to flang-new -fc1 avoids unnecessary complexity in the frontend driver.

-menable-infs is removed from the frontend-forwarding test because if all of the fast-math component flags are present, these will be resolved into the fast-math flag. Instead -menable-infs is tested in the fast-math test.

-fno-fast-math will not make FPContract less strict: -ffp-contract=off -fno-fast-math will not reset to -ffp-contract=fast (the default).

[awarzynski]

Hi Tom, thanks for working on this! I'm not that familiar with the semantics of -Ofast (and the related flags), but everything else looks good. Some comments below and inline.

-Ofast is defined as -O3 -ffast-math

In the compiler driver,flang-new, or the frontend driver, flang-new -fc1?

This definition matches the meaning of the "fast" attribute in LLVM IR: https://llvm.org/docs/LangRef.html#fast-math-flags

[nit] But there's not guarantee just yet that -ofast will result in the attributes being added? In fact, why not add a test to verify that?

Another related question - are the semantics that you are implementing here identical to Clang? Either way, could you document?

clang -cc1 accepts -Ofast. I did not add it to flang -fc1 because this seems redundant because the compiler driver will always resolve -Ofast into -O3 -ffast-math (I added a test for this).

I don't quite follow - Clang users can run clang -cc1 -Ofast. What's the equivalent in Flang?

-fstack-arrays should also be enabled with -Ofast but that flang is not yet implemented and so has been left for a future patch.

Why not add a test?

flang/lib/Frontend/CompilerInvocation.cpp

733

Why is this diagnostic needed? And is it tested?

flang/test/Driver/fast_math.f90

8–9

Similar suggestion for other tests in this file (whenever it makes sense)

[tblah]

In D138675#3951836, @awarzynski wrote:

Hi Tom, thanks for working on this! I'm not that familiar with the semantics of -Ofast (and the related flags), but everything else looks good. Some comments below and inline.

-Ofast is defined as -O3 -ffast-math

In the compiler driver,flang-new, or the frontend driver, flang-new -fc1?

In the compiler driver. flang-new -fc1 does not accept -Ofast.

This definition matches the meaning of the "fast" attribute in LLVM IR: https://llvm.org/docs/LangRef.html#fast-math-flags

[nit] But there's not guarantee just yet that -ofast will result in the attributes being added? In fact, why not add a test to verify that?

There is a test checking that -Ofast does generate -ffast-math (flang/test/Driver/fast_math.f90), and that -ffast-math leads to FIR with the fast attribute (flang/test/Lower/fast-math-arithmetic.f90).

Another related question - are the semantics that you are implementing here identical to Clang? Either way, could you document?

clang -cc1 accepts -Ofast. I did not add it to flang -fc1 because this seems redundant because the compiler driver will always resolve -Ofast into -O3 -ffast-math (I added a test for this).

I don't quite follow - Clang users can run clang -cc1 -Ofast. What's the equivalent in Flang?

I guess the equivalent in Flang would be flang-new -cc1 -O3 -ffast-math. As I understand it, flang-new -fc1 is not intended as an interface for users.

-fstack-arrays should also be enabled with -Ofast but that flang is not yet implemented and so has been left for a future patch.

Why not add a test?

Because that flag has not been added to Flang yet

[tblah]

flang/lib/Frontend/CompilerInvocation.cpp

733

@vzakhari didn't want to remove these warnings when support was first added because the intrinsics simplification pass still ignored fastmath (https://reviews.llvm.org/D137391#3908424).

There has been work on the intrinsics since then in https://reviews.llvm.org/D138048 but I don't know if this is complete. @vzakhari would you like me to remove these warnings for the implemented flags now?

The test is at the end of flang/test/Driver/fast_math.f90

[tblah]

Update tests to check that forwarded flags are on the same line as -fc1

[awarzynski]

In D138675#3952712, @tblah wrote:

In D138675#3951836, @awarzynski wrote:

Hi Tom, thanks for working on this! I'm not that familiar with the semantics of -Ofast (and the related flags), but everything else looks good. Some comments below and inline.

-Ofast is defined as -O3 -ffast-math

In the compiler driver,flang-new, or the frontend driver, flang-new -fc1?

In the compiler driver. flang-new -fc1 does not accept -Ofast.

So, -Ofast in flang-new is translated as flang-new -fc1 -O3 -ffast-math? Could you clarify this in the summary? Otherwise, there is no documentation for this.

This definition matches the meaning of the "fast" attribute in LLVM IR: https://llvm.org/docs/LangRef.html#fast-math-flags

[nit] But there's not guarantee just yet that -ofast will result in the attributes being added? In fact, why not add a test to verify that?

There is a test checking that -Ofast does generate -ffast-math (flang/test/Driver/fast_math.f90), and that -ffast-math leads to FIR with the fast attribute (flang/test/Lower/fast-math-arithmetic.f90).

Worth adding a note in the summary.

Another related question - are the semantics that you are implementing here identical to Clang? Either way, could you document?

clang -cc1 accepts -Ofast. I did not add it to flang -fc1 because this seems redundant because the compiler driver will always resolve -Ofast into -O3 -ffast-math (I added a test for this).

So what does clang -cc1 -Ofast mean? And how come we don't need this in flang-new -fc1?

I don't quite follow - Clang users can run clang -cc1 -Ofast. What's the equivalent in Flang?

I guess the equivalent in Flang would be flang-new -cc1 -O3 -ffast-math. As I understand it, flang-new -fc1 is not intended as an interface for users.

I would recommend following the principle of least surprise and just follow what Clang does.

-fstack-arrays should also be enabled with -Ofast but that flang is not yet implemented and so has been left for a future patch.

Why not add a test?

Because that flag has not been added to Flang yet

I don't follow - you can still write a test for a flag that's not supported.

[vzakhari]

flang/include/flang/Common/MathOptionsBase.def

42 ↗

(On Diff #478159)

I do not think this is needed here. In LLVM fast is just a decoration for "all-bits-set", so I do not see why we would need a separate bit for this. Moreover, if you add it we will have to guarantee and verify consistency (i.e. that FastMath == all-above-bits-are-set).

flang/lib/Frontend/CompilerInvocation.cpp

733

Yes, I think we can remove the warnings now.

[tblah]

Update to remove the warning that -ffast-math is unimplemented and do not store fast-math status independently of the flags it contains.

[tblah]

flang/lib/Frontend/CompilerInvocation.cpp

733

I have removed the rest of the warnings in https://reviews.llvm.org/D138907

[tblah]

Add test for -fstack-arrays usage warning

[awarzynski]

Thanks for the updates. A few more nits from me. LGTM and ready to land once addressed.

@vzakhari ?

clang/lib/Driver/ToolChains/Flang.cpp

197–200

So this means that:

• --ffp-contract=off -fno-fast-math --> FPContract = "off"
• --fno-fast-math --> FPContract = ""?

I would reduce this comment to that ^^^. In particular, what if this is no longer true: "// FPContract = fast is the default anyway" ?

Also, could you mention this in the summary? This a design decision worth documenting.

213

Is this comment needed? Would other flags do you have in mind?

[tblah]

Updated comment, updated patch context, fixed unused variable

[tblah]

Updating patch add documentation of the meaning of -Ofast, -ffast-math, and comparison with gfortran and nvfortran.

Also added a test to ensure that crtfastmath.o is linked when -ffast-math is specified to the compiler driver.

[tblah]

Update patch to skip the test for linking to crtfastmath.o on Windows as there doesn't seem to be a Windows equivalent of crtfastmath.o.

[tblah]

Failed Tests (5):

Flang :: Fir/boxproc.fir
Flang :: Fir/target-rewrite-arg-position.fir
Flang :: Fir/target-rewrite-boxchar.fir
Flang :: Fir/target-rewrite-char-proc.fir
Flang :: Fir/target.fir

I am getting the same without my patch (upstream ref 38e87e8af47a)
 

[kiranchandramohan]

LGTM. Thanks for addressing all the comments and comparing the option with other Fortran compilers. Please wait a day before submitting in case there are other comments.

flang/docs/FlangDriver.md

552

Nit: Add a link to what these options mean. Possibly say that these options are required to set the various fast math attributes on the LLVM IR instructions as listed in the following link.
https://llvm.org/docs/LangRef.html#fast-math-flags

588

Nit: You can say that the O3 pipeline has passes that perform transformations like inlining, vectorisation, unrolling etc. Additionally the GVN, LICM passes performs redundancy elimination like Mpre, Mlre.

590

Nit: I guess the plan is not to include this flag even when fomit-frame-pointer support is added.

[vzakhari]

Thank you for the additional changes, @tblah! LGTM

[tblah]

Update with nitpicks to documentation (mostly to re-run Windows CI)

[clementval]

I have the Driver/fast_math.f90 failing on my side.

llvm-project/flang/test/Driver/fast_math.f90:63:14: error: CHECK-CRT: expected string not found in input
! CHECK-CRT: crtbeginS.o

I have a crtbegin.o in the link line but not a crtbeginS.o

[tblah]

In D138675#3987836, @clementval wrote:

I have the Driver/fast_math.f90 failing on my side.

llvm-project/flang/test/Driver/fast_math.f90:63:14: error: CHECK-CRT: expected string not found in input
! CHECK-CRT: crtbeginS.o

I have a crtbegin.o in the link line but not a crtbeginS.o

Thanks for the heads up. This should be fixed by https://github.com/llvm/llvm-project/commit/bea45027b43ec61a3efc4b487ceca2da25504432

[clementval]

In D138675#3988110, @tblah wrote:

In D138675#3987836, @clementval wrote:

I have the Driver/fast_math.f90 failing on my side.

llvm-project/flang/test/Driver/fast_math.f90:63:14: error: CHECK-CRT: expected string not found in input
! CHECK-CRT: crtbeginS.o

I have a crtbegin.o in the link line but not a crtbeginS.o

Thanks for the heads up. This should be fixed by https://github.com/llvm/llvm-project/commit/bea45027b43ec61a3efc4b487ceca2da25504432

Thanks for the fix. crtendS.o has the same issue.

[amyk]

Hi,

The ppc64le-flang-rhel-clang bot is also experiencing some failures with the Driver/fast_math.f90 test case. In particular, crtfastmath.o is not found at all for CHECK-CRT as we can see in the failure details.

llvm-project/flang/test/Driver/fast_math.f90:64:19: error: CHECK-CRT-SAME: expected string not found in input
! CHECK-CRT-SAME: crtfastmath.o

The output for the CHECK-CRT line on the bot looks like the following:

"/home/buildbots/ppc64le-flang-rhel-test/ppc64le-flang-rhel-clang-build/build/bin/flang-new" "-fc1" "-triple" "powerpc64le-unknown-linux-gnu" "-emit-obj" "-mrelocation-model" "pic" "-pic-level" "2" "-pic-is-pie" "-ffast-math" "-target-cpu" "ppc64le" "-o" "/tmp/lit-tmp-x_okwzug/fast_math-e9ab49.o" "-x" "f95-cpp-input" "/home/buildbots/ppc64le-flang-rhel-test/ppc64le-flang-rhel-clang-build/llvm-project/flang/test/Driver/fast_math.f90" 

"/usr/bin/ld" "-pie" "--hash-style=gnu" "--eh-frame-hdr" "-m" "elf64lppc" "-dynamic-linker" "/lib64/ld64.so.2" "-o" "/home/buildbots/ppc64le-flang-rhel-test/ppc64le-flang-rhel-clang-build/build/tools/flang/test/Driver/Output/fast_math.f90.tmp" "/usr/lib/gcc/ppc64le-redhat-linux/8/../../../../lib64/Scrt1.o" "/usr/lib/gcc/ppc64le-redhat-linux/8/../../../../lib64/crti.o" "/usr/lib/gcc/ppc64le-redhat-linux/8/crtbeginS.o" "-L/usr/lib/gcc/ppc64le-redhat-linux/8" "-L/usr/lib/gcc/ppc64le-redhat-linux/8/../../../../lib64" "-L/lib/../lib64" "-L/usr/lib/../lib64" "-L/lib" "-L/usr/lib" "/tmp/lit-tmp-x_okwzug/fast_math-e9ab49.o" "-lFortran_main" "-lFortranRuntime" "-lFortranDecimal" "-lm" "-lgcc" "--as-needed" "-lgcc_s" "--no-as-needed" "-lc" "-lgcc" "--as-needed" "-lgcc_s" "--no-as-needed" "/usr/lib/gcc/ppc64le-redhat-linux/8/crtendS.o" "/usr/lib/gcc/ppc64le-redhat-linux/8/../../../../lib64/crtn.o"

Would it be possible to follow up with a fix for this test case?

[tblah]

In D138675#3989403, @amyk wrote:

Hi,

The ppc64le-flang-rhel-clang bot is also experiencing some failures with the Driver/fast_math.f90 test case. In particular, crtfastmath.o is not found at all for CHECK-CRT as we can see in the failure details.

llvm-project/flang/test/Driver/fast_math.f90:64:19: error: CHECK-CRT-SAME: expected string not found in input
! CHECK-CRT-SAME: crtfastmath.o

The output for the CHECK-CRT line on the bot looks like the following:

"/home/buildbots/ppc64le-flang-rhel-test/ppc64le-flang-rhel-clang-build/build/bin/flang-new" "-fc1" "-triple" "powerpc64le-unknown-linux-gnu" "-emit-obj" "-mrelocation-model" "pic" "-pic-level" "2" "-pic-is-pie" "-ffast-math" "-target-cpu" "ppc64le" "-o" "/tmp/lit-tmp-x_okwzug/fast_math-e9ab49.o" "-x" "f95-cpp-input" "/home/buildbots/ppc64le-flang-rhel-test/ppc64le-flang-rhel-clang-build/llvm-project/flang/test/Driver/fast_math.f90" 

"/usr/bin/ld" "-pie" "--hash-style=gnu" "--eh-frame-hdr" "-m" "elf64lppc" "-dynamic-linker" "/lib64/ld64.so.2" "-o" "/home/buildbots/ppc64le-flang-rhel-test/ppc64le-flang-rhel-clang-build/build/tools/flang/test/Driver/Output/fast_math.f90.tmp" "/usr/lib/gcc/ppc64le-redhat-linux/8/../../../../lib64/Scrt1.o" "/usr/lib/gcc/ppc64le-redhat-linux/8/../../../../lib64/crti.o" "/usr/lib/gcc/ppc64le-redhat-linux/8/crtbeginS.o" "-L/usr/lib/gcc/ppc64le-redhat-linux/8" "-L/usr/lib/gcc/ppc64le-redhat-linux/8/../../../../lib64" "-L/lib/../lib64" "-L/usr/lib/../lib64" "-L/lib" "-L/usr/lib" "/tmp/lit-tmp-x_okwzug/fast_math-e9ab49.o" "-lFortran_main" "-lFortranRuntime" "-lFortranDecimal" "-lm" "-lgcc" "--as-needed" "-lgcc_s" "--no-as-needed" "-lc" "-lgcc" "--as-needed" "-lgcc_s" "--no-as-needed" "/usr/lib/gcc/ppc64le-redhat-linux/8/crtendS.o" "/usr/lib/gcc/ppc64le-redhat-linux/8/../../../../lib64/crtn.o"

Would it be possible to follow up with a fix for this test case?

Thanks for letting me know. I thought I had already disabled that test on powerpc (in https://github.com/llvm/llvm-project/commit/20cd3153f3775fcdc1eeeb54062849eead51e24a) but apparently it didn't work. I don't have a powerpc system available to test this. I have pushed a second attempt at https://github.com/llvm/llvm-project/commit/6442b4da4e7018e8f264965768b9e4fdee393c8f. Please let me know if that works.

[amyk]

In D138675#3989634, @tblah wrote:

In D138675#3989403, @amyk wrote:

Hi,

The ppc64le-flang-rhel-clang bot is also experiencing some failures with the Driver/fast_math.f90 test case. In particular, crtfastmath.o is not found at all for CHECK-CRT as we can see in the failure details.

llvm-project/flang/test/Driver/fast_math.f90:64:19: error: CHECK-CRT-SAME: expected string not found in input
! CHECK-CRT-SAME: crtfastmath.o

The output for the CHECK-CRT line on the bot looks like the following:

"/home/buildbots/ppc64le-flang-rhel-test/ppc64le-flang-rhel-clang-build/build/bin/flang-new" "-fc1" "-triple" "powerpc64le-unknown-linux-gnu" "-emit-obj" "-mrelocation-model" "pic" "-pic-level" "2" "-pic-is-pie" "-ffast-math" "-target-cpu" "ppc64le" "-o" "/tmp/lit-tmp-x_okwzug/fast_math-e9ab49.o" "-x" "f95-cpp-input" "/home/buildbots/ppc64le-flang-rhel-test/ppc64le-flang-rhel-clang-build/llvm-project/flang/test/Driver/fast_math.f90" 

"/usr/bin/ld" "-pie" "--hash-style=gnu" "--eh-frame-hdr" "-m" "elf64lppc" "-dynamic-linker" "/lib64/ld64.so.2" "-o" "/home/buildbots/ppc64le-flang-rhel-test/ppc64le-flang-rhel-clang-build/build/tools/flang/test/Driver/Output/fast_math.f90.tmp" "/usr/lib/gcc/ppc64le-redhat-linux/8/../../../../lib64/Scrt1.o" "/usr/lib/gcc/ppc64le-redhat-linux/8/../../../../lib64/crti.o" "/usr/lib/gcc/ppc64le-redhat-linux/8/crtbeginS.o" "-L/usr/lib/gcc/ppc64le-redhat-linux/8" "-L/usr/lib/gcc/ppc64le-redhat-linux/8/../../../../lib64" "-L/lib/../lib64" "-L/usr/lib/../lib64" "-L/lib" "-L/usr/lib" "/tmp/lit-tmp-x_okwzug/fast_math-e9ab49.o" "-lFortran_main" "-lFortranRuntime" "-lFortranDecimal" "-lm" "-lgcc" "--as-needed" "-lgcc_s" "--no-as-needed" "-lc" "-lgcc" "--as-needed" "-lgcc_s" "--no-as-needed" "/usr/lib/gcc/ppc64le-redhat-linux/8/crtendS.o" "/usr/lib/gcc/ppc64le-redhat-linux/8/../../../../lib64/crtn.o"

Would it be possible to follow up with a fix for this test case?

Thanks for letting me know. I thought I had already disabled that test on powerpc (in https://github.com/llvm/llvm-project/commit/20cd3153f3775fcdc1eeeb54062849eead51e24a) but apparently it didn't work. I don't have a powerpc system available to test this. I have pushed a second attempt at https://github.com/llvm/llvm-project/commit/6442b4da4e7018e8f264965768b9e4fdee393c8f. Please let me know if that works.

Thanks for the follow up patch! I tested the patch locally and also saw the buildbot results, and it doesn't appear like the follow up patch marked powerpc as unsupported as the error still persists (https://lab.llvm.org/buildbot/#/builders/21/builds/57850).
I was playing around with the test case locally and what appears to work is doing something like:

! UNSUPPORTED: powerpc-registered-target

[tblah]

In D138675#3989826, @amyk wrote:

Thanks for the follow up patch! I tested the patch locally and also saw the buildbot results, and it doesn't appear like the follow up patch marked powerpc as unsupported as the error still persists (https://lab.llvm.org/buildbot/#/builders/21/builds/57850).
I was playing around with the test case locally and what appears to work is doing something like:

! UNSUPPORTED: powerpc-registered-target

Thanks https://github.com/llvm/llvm-project/commit/9d86f2dc4f1d2e4e1a991be82384bbdb310f0618

[amyk]

In D138675#3989997, @tblah wrote:

In D138675#3989826, @amyk wrote:

Thanks for the follow up patch! I tested the patch locally and also saw the buildbot results, and it doesn't appear like the follow up patch marked powerpc as unsupported as the error still persists (https://lab.llvm.org/buildbot/#/builders/21/builds/57850).
I was playing around with the test case locally and what appears to work is doing something like:

! UNSUPPORTED: powerpc-registered-target

Thanks https://github.com/llvm/llvm-project/commit/9d86f2dc4f1d2e4e1a991be82384bbdb310f0618

Thanks for following up! The bot is now green (https://lab.llvm.org/buildbot/#/builders/21/builds/57857).

[probinson]

See D139967 for why UNSUPPORTED: powerpc didn't work. That patch will put it back, and also update the lit config so the check will work now.

[tblah]

In D138675#3995282, @probinson wrote:

See D139967 for why UNSUPPORTED: powerpc didn't work. That patch will put it back, and also update the lit config so the check will work now.

Thanks for the update!

[amyk]

In D138675#3995282, @probinson wrote:

See D139967 for why UNSUPPORTED: powerpc didn't work. That patch will put it back, and also update the lit config so the check will work now.

Thank you for following up with this. I appreciate it!

[mnadeem]

flang/test/Driver/fast_math.f90

59–62

I think this test should use an explicit sysroot like the clang test does. It fails on my system (x86 linux, llvm built for arm target) because it cant find the crtfastmath.o library.

// file linux-ld.c

RUN: --sysroot=%S/Inputs/basic_linux_tree 2>&1 \
RUN: | FileCheck --check-prefix=CHECK-CRTFASTMATH %s

[tblah]

flang/test/Driver/fast_math.f90

59–62

Thanks for the suggestion https://reviews.llvm.org/D142807

clang/include/clang/Driver/Options.td

 def O0 : Flag<["-"], "O0">, Group<O_Group>, Flags<[CC1Option, FC1Option, HelpHidden]>;
 def O4 : Flag<["-"], "O4">, Group<O_Group>, Flags<[CC1Option, FC1Option, HelpHidden]>;
 def ObjCXX : Flag<["-"], "ObjC++">, Flags<[NoXarchOption]>,
   HelpText<"Treat source input files as Objective-C++ inputs">;
 def ObjC : Flag<["-"], "ObjC">, Flags<[NoXarchOption]>,
   HelpText<"Treat source input files as Objective-C inputs">;
 def O : Joined<["-"], "O">, Group<O_Group>, Flags<[CC1Option,FC1Option]>;
 def O_flag : Flag<["-"], "O">, Flags<[CC1Option,FC1Option]>, Alias<O>, AliasArgs<["1"]>;
-def Ofast : Joined<["-"], "Ofast">, Group<O_Group>, Flags<[CC1Option]>;
+def Ofast : Joined<["-"], "Ofast">, Group<O_Group>, Flags<[CC1Option, FlangOption]>;
 def P : Flag<["-"], "P">, Flags<[CC1Option,FlangOption,FC1Option]>, Group<Preprocessor_Group>,
   HelpText<"Disable linemarker output in -E mode">,
   MarshallingInfoNegativeFlag<PreprocessorOutputOpts<"ShowLineMarkers">>;
 def Qy : Flag<["-"], "Qy">, Flags<[CC1Option]>,
   HelpText<"Emit metadata containing compiler name and version">;
 def Qn : Flag<["-"], "Qn">, Flags<[CC1Option]>,
   HelpText<"Do not emit metadata containing compiler name and version">;
 def : Flag<["-"], "fident">, Group<f_Group>, Alias<Qy>,
def ffp_model_EQ : Joined<["-"], "ffp-model=">, Group<f_Group>, Flags<[NoXarchOption]>,
   HelpText<"Controls the semantics of floating-point calculations.">;
 def ffp_exception_behavior_EQ : Joined<["-"], "ffp-exception-behavior=">, Group<f_Group>, Flags<[CC1Option]>,
   HelpText<"Specifies the exception behavior of floating-point operations.">,
   Values<"ignore,maytrap,strict">, NormalizedValuesScope<"LangOptions">,
   NormalizedValues<["FPE_Ignore", "FPE_MayTrap", "FPE_Strict"]>,
   MarshallingInfoEnum<LangOpts<"FPExceptionMode">, "FPE_Default">;
 defm fast_math : BoolFOption<"fast-math",
   LangOpts<"FastMath">, DefaultFalse,
-  PosFlag<SetTrue, [CC1Option], "Allow aggressive, lossy floating-point optimizations",
+  PosFlag<SetTrue, [CC1Option, FC1Option, FlangOption], "Allow aggressive, lossy floating-point optimizations",
           [cl_fast_relaxed_math.KeyPath]>,
   NegFlag<SetFalse>>;
 defm math_errno : BoolFOption<"math-errno",
   LangOpts<"MathErrno">, DefaultFalse,
   PosFlag<SetTrue, [CC1Option], "Require math functions to indicate errors by setting errno">,
   NegFlag<SetFalse>>,
   ShouldParseIf<!strconcat("!", open_cl.KeyPath)>;
 def fextend_args_EQ : Joined<["-"], "fextend-arguments=">, Group<f_Group>,

clang/lib/Driver/ToolChains/Flang.cpp

case options::OPT_fno_associative_math:
       AssociativeMath = false;
       break;
     case options::OPT_freciprocal_math:
       ReciprocalMath = true;
       break;
     case options::OPT_fno_reciprocal_math:
       ReciprocalMath = false;
       break;
+    case options::OPT_Ofast:
+      [[fallthrough]];
+    case options::OPT_ffast_math:
+      HonorINFs = false;
+      HonorNaNs = false;
+      AssociativeMath = true;
+      ReciprocalMath = true;
+      ApproxFunc = true;
+      SignedZeros = false;
+      FPContract = "fast";
+      break;
+    case options::OPT_fno_fast_math:
+      HonorINFs = true;
+      HonorNaNs = true;
+      AssociativeMath = false;
+      ReciprocalMath = false;
+      ApproxFunc = false;
+      SignedZeros = true;
+      // -fno-fast-math should undo -ffast-math so I return FPContract to the
+      // default. It is important to check it is "fast" (the default) so that
+      // --ffp-contract=off -fno-fast-math --> -ffp-contract=off
+      if (FPContract == "fast")
+        FPContract = "";
+      break;
     }
 
     // If we handled this option claim it
     A->claim();
   }
 
+  if (!HonorINFs && !HonorNaNs && AssociativeMath && ReciprocalMath &&
+      ApproxFunc && !SignedZeros &&
+      (FPContract == "fast" || FPContract == "")) {
+    CmdArgs.push_back("-ffast-math");
+    return;
+  }
+
   if (!FPContract.empty())
     CmdArgs.push_back(Args.MakeArgString("-ffp-contract=" + FPContract));
 
   if (!HonorINFs)
     CmdArgs.push_back("-menable-no-infs");
 
   if (!HonorNaNs)
     CmdArgs.push_back("-menable-no-nans");
for (const Arg *A : Args.filtered(options::OPT_mmlir)) {
     A->render(Args, CmdArgs);
   }
 
   // Optimization level for CodeGen.
   if (const Arg *A = Args.getLastArg(options::OPT_O_Group)) {
     if (A->getOption().matches(options::OPT_O4)) {
       CmdArgs.push_back("-O3");
       D.Diag(diag::warn_O4_is_O3);
+    } else if (A->getOption().matches(options::OPT_Ofast)) {
+      CmdArgs.push_back("-O3");
     } else {
       A->render(Args, CmdArgs);
     }
   }
 
   if (Output.isFilename()) {
     CmdArgs.push_back("-o");
     CmdArgs.push_back(Output.getFilename());

flang/docs/FlangDriver.md

 statically. Setting `-DLLVM_${NAME}_LINK_INTO_TOOLS` to `ON` in the cmake
 command turns the project into a statically linked extension. An example would
 be Polly, e.g., using `-DLLVM_POLLY_LINK_INTO_TOOLS=ON` would link Polly passes
 into `flang-new` as built-in middle-end passes.
 
 See the
 [`WritingAnLLVMNewPMPass`](https://llvm.org/docs/WritingAnLLVMNewPMPass.html#id9)
 documentation for more details.
+
+## Ofast and Fast Math
+`-Ofast` in Flang means `-O3 -ffast-math`. `-fstack-arrays` will be added to
+`-Ofast` in the future (https://github.com/llvm/llvm-project/issues/59231).
+
+`-ffast-math` means the following:
+ - `-fno-honor-infinities`
+ - `-fno-honor-nans`
+ - `-fassociative-math`
+ - `-freciprocal-math`
+ - `-fapprox-func`
+ - `-fno-signed-zeros`
+ - `-ffp-contract=fast`
+
+These correspond to LLVM IR Fast Math attributes:
+https://llvm.org/docs/LangRef.html#fast-math-flags
+
+When `-ffast-math` is specified, any linker steps generated by the compiler
+driver will also link to `crtfastmath.o`, which adds a static constructor
+that sets the FTZ/DAZ bits in MXCSR, affecting not only the current only the
+current compilation unit but all static and shared libraries included in the
+program. Setting these bits causes denormal floating point numbers to be flushed
+to zero.
+
+### Comparison with GCC/GFortran
+GCC/GFortran translate `-Ofast` to
+`-O3 -ffast-math -fstack-arrays -fno-semantic-interposition`. `-fstack-arrays`
+is TODO for Flang.
+`-fno-semantic-interposition` is not used because clang does not enable this as
+part of `-Ofast` as the default behaviour is similar.
+
+GCC/GFortran has a wider definition of `-ffast-math`: also including
+`-fno-trapping-math`,  `-fno-rounding-math`, and  `-fsignaling-nans`; these
+aren't included in Flang because Flang currently has no support for strict
+floating point and so always acts as though these flags were specified.
+
+GCC/GFortran will also set flush-to-zero mode: linking `crtfastmath.o`, the same
+as Flang.
+
+### Comparison with nvfortran
+nvfortran defines `-fast` as
+`-O2 -Munroll=c:1 -Mnoframe -Mlre -Mpre -Mvect=simd -Mcache_align -Mflushz -Mvect`.
+ - `-O2 -Munroll=c:1 -Mlre -Mautoinline -Mpre -Mvect-simd` affect code
+   optimization. `flang -O3` should enable all optimizations for execution time,
+   similarly to `clang -O3`. The `-O3` pipeline has passes that perform
+   transformations like inlining, vectorisation, unrolling, etc. Additionally,
+   the GVN and LICM passes perform redundancy elimination like `Mpre` and `Mlre`
+ - `-Mnoframe`: the equivalent flag would be `-fomit-frame-pointer`. This flag
+   is not yet supported in Flang and so Flang follows GFortran in not including
+   this in `-Ofast`. There is no plan to include this flag as part of `-Ofast`.
+ - `-Mcache_align`: there is no equivalent flag in Flang or Clang.
+ - `-Mflushz`: flush-to-zero mode - when `-ffast-math` is specified, Flang will
+   link to `crtfastmath.o` to ensure denormal numbers are flushed to zero.

flang/lib/Frontend/CompilerInvocation.cpp

static bool parseFloatingPointArgs(CompilerInvocation &invoc,
   if (args.getLastArg(clang::driver::options::OPT_mreassociate)) {
     opts.AssociativeMath = true;
   }
 
   if (args.getLastArg(clang::driver::options::OPT_freciprocal_math)) {
     opts.ReciprocalMath = true;
   }
 
+  if (args.getLastArg(clang::driver::options::OPT_ffast_math)) {
+    opts.NoHonorInfs = true;
+    opts.NoHonorNaNs = true;
+    opts.AssociativeMath = true;
+    opts.ReciprocalMath = true;
+    opts.ApproxFunc = true;
+    opts.NoSignedZeros = true;
+    opts.setFPContractMode(LangOptions::FPM_Fast);
+  }
+
   return true;
 }
 
 bool CompilerInvocation::createFromArgs(
     CompilerInvocation &res, llvm::ArrayRef<const char *> commandLineArgs,
     clang::DiagnosticsEngine &diags) {
 
   bool success = true;

flang/test/Driver/driver-help-hidden.f90

 ! CHECK-NEXT: Enable the old style PARAMETER statement
 ! CHECK-NEXT: -fapprox-func          Allow certain math function calls to be replaced with an approximately equivalent calculation
 ! CHECK-NEXT: -fbackslash            Specify that backslash in string introduces an escape character
 ! CHECK-NEXT: -fcolor-diagnostics    Enable colors in diagnostics
 ! CHECK-NEXT: -fconvert=<value>      Set endian conversion of data for unformatted files
 ! CHECK-NEXT: -fdefault-double-8     Set the default double precision kind to an 8 byte wide type
 ! CHECK-NEXT: -fdefault-integer-8    Set the default integer kind to an 8 byte wide type
 ! CHECK-NEXT: -fdefault-real-8       Set the default real kind to an 8 byte wide type
+! CHECK-NEXT: -ffast-math            Allow aggressive, lossy floating-point optimizations
 ! CHECK-NEXT: -ffixed-form           Process source files in fixed form
 ! CHECK-NEXT: -ffixed-line-length=<value>
 ! CHECK-NEXT: Use <value> as character line width in fixed mode
 ! CHECK-NEXT: -ffp-contract=<value> Form fused FP ops (e.g. FMAs)
 ! CHECK-NEXT: -ffree-form            Process source files in free form
 ! CHECK-NEXT: -fimplicit-none        No implicit typing allowed unless overridden by IMPLICIT statements
 ! CHECK-NEXT: -finput-charset=<value> Specify the default character set for source files
 ! CHECK-NEXT: -fintrinsic-modules-path <dir>

flang/test/Driver/driver-help.f90

 ! HELP-NEXT: Enable the old style PARAMETER statement
 ! HELP-NEXT: -fapprox-func          Allow certain math function calls to be replaced with an approximately equivalent calculation
 ! HELP-NEXT: -fbackslash            Specify that backslash in string introduces an escape character
 ! HELP-NEXT: -fcolor-diagnostics    Enable colors in diagnostics
 ! HELP-NEXT: -fconvert=<value>      Set endian conversion of data for unformatted files
 ! HELP-NEXT: -fdefault-double-8     Set the default double precision kind to an 8 byte wide type
 ! HELP-NEXT: -fdefault-integer-8    Set the default integer kind to an 8 byte wide type
 ! HELP-NEXT: -fdefault-real-8       Set the default real kind to an 8 byte wide type
+! HELP-NEXT: -ffast-math            Allow aggressive, lossy floating-point optimizations
 ! HELP-NEXT: -ffixed-form           Process source files in fixed form
 ! HELP-NEXT: -ffixed-line-length=<value>
 ! HELP-NEXT: Use <value> as character line width in fixed mode
 ! HELP-NEXT: -ffp-contract=<value> Form fused FP ops (e.g. FMAs)
 ! HELP-NEXT: -ffree-form            Process source files in free form
 ! HELP-NEXT: -fimplicit-none        No implicit typing allowed unless overridden by IMPLICIT statements
 ! HELP-NEXT: -finput-charset=<value> Specify the default character set for source files
 ! HELP-NEXT: -fintrinsic-modules-path <dir>

 ! HELP-FC1-NEXT: -fdebug-pre-fir-tree    Dump the pre-FIR tree
 ! HELP-FC1-NEXT: -fdebug-unparse-no-sema Unparse and stop (skips the semantic checks)
 ! HELP-FC1-NEXT: -fdebug-unparse-with-symbols
 ! HELP-FC1-NEXT:                        Unparse and stop.
 ! HELP-FC1-NEXT: -fdebug-unparse        Unparse and stop.
 ! HELP-FC1-NEXT: -fdefault-double-8  Set the default double precision kind to an 8 byte wide type
 ! HELP-FC1-NEXT: -fdefault-integer-8 Set the default integer kind to an 8 byte wide type
 ! HELP-FC1-NEXT: -fdefault-real-8    Set the default real kind to an 8 byte wide type
+! HELP-FC1-NEXT: -ffast-math            Allow aggressive, lossy floating-point optimizations
 ! HELP-FC1-NEXT: -ffixed-form           Process source files in fixed form
 ! HELP-FC1-NEXT: -ffixed-line-length=<value>
 ! HELP-FC1-NEXT: Use <value> as character line width in fixed mode
 ! HELP-FC1-NEXT: -ffp-contract=<value> Form fused FP ops (e.g. FMAs)
 ! HELP-FC1-NEXT: -ffree-form            Process source files in free form
 ! HELP-FC1-NEXT: -fget-definition <value> <value> <value>
 ! HELP-FC1-NEXT:                        Get the symbol definition from <line> <start-column> <end-column>
 ! HELP-FC1-NEXT: -fget-symbols-sources   Dump symbols and their source code locations

flang/test/Driver/fast_math.f90

+! Test for correct forwarding of fast-math flags from the compiler driver to the
+! frontend driver
+
+! -Ofast => -ffast-math -O3
+! RUN: %flang -Ofast -fsyntax-only -### %s -o %t 2>&1 \
+! RUN:     | FileCheck --check-prefix=CHECK-OFAST %s
+! CHECK-OFAST: -fc1
+! CHECK-OFAST-SAME: -ffast-math
+! CHECK-OFAST-SAME: -O3
+
+! TODO: update once -fstack-arays is added
+! RUN: %flang -fstack-arrays -fsyntax-only %s -o %t 2>&1 \
+! RUN:     | FileCheck --check-prefix=CHECK-STACK-ARRAYS %s
+! CHECK-STACK-ARRAYS: warning: argument unused during compilation: '-fstack-arrays'
+
+! -Ofast -fno-fast-math => -O3
+! RUN: %flang -Ofast -fno-fast-math -fsyntax-only -### %s -o %t 2>&1 \
+! RUN:     | FileCheck --check-prefix=CHECK-OFAST-NO-FAST %s
+! CHECK-OFAST-NO-FAST: -fc1
+! CHECK-OFAST-NO-FAST-NOT: -ffast-math
+! CHECK-OFAST-NO-FAST-SAME: -O3
+
+! -ffast-math => -ffast-math
+! RUN: %flang -ffast-math -fsyntax-only -### %s -o %t 2>&1 \
+! RUN:     | FileCheck --check-prefix=CHECK-FFAST %s
+! CHECK-FFAST: -fc1
+! CHECK-FFAST-SAME: -ffast-math
+
+! (component flags) => -ffast-math
+! RUN: %flang -fsyntax-only -### %s -o %t \
+! RUN:     -fno-honor-infinities \
+! RUN:     -fno-honor-nans \
+! RUN:     -fassociative-math \
+! RUN:     -freciprocal-math \
+! RUN:     -fapprox-func \
+! RUN:     -fno-signed-zeros \
+! RUN:     -ffp-contract=fast \
+! RUN:     2>&1 | FileCheck --check-prefix=CHECK-FROM-COMPS %s
+! CHECK-FROM-COMPS: -fc1
+! CHECK-FROM-COMPS-SAME: -ffast-math
+
+! -ffast-math (followed by an alteration) => (component flags)
+! RUN: %flang -ffast-math -fhonor-infinities -fsyntax-only -### %s -o %t 2>&1 \
+! RUN:     | FileCheck --check-prefix=CHECK-TO-COMPS %s
+! CHECK-TO-COMPS: -fc1
+! CHECK-TO-COMPS-SAME: -ffp-contract=fast
+! CHECK-TO-COMPS-SAME: -menable-no-nans
+! CHECK-TO-COMPS-SAME: -fapprox-func
+! CHECK-TO-COMPS-SAME: -fno-signed-zeros
+! CHECK-TO-COMPS-SAME: -mreassociate
+! CHECK-TO-COMPS-SAME: -freciprocal-math
+
+! Check that -fno-fast-math doesn't clobber -ffp-contract
+! RUN: %flang -ffp-contract=off -fno-fast-math -fsyntax-only -### %s -o %t 2>&1 \
+! RUN:     | FileCheck --check-prefix=CHECK-CONTRACT %s
+! CHECK-CONTRACT: -fc1
+! CHECK-CONTRACT-SAME: -ffp-contract=off
+
+! Check that -ffast-math causes us to link to crtfastmath.o
+! UNSUPPORTED: system-windows
+! RUN: %flang -ffast-math -### %s -o %t 2>&1 \
+! RUN:     | FileCheck --check-prefix=CHECK-CRT %s
+! CHECK-CRT: crtbeginS.o
+! CHECK-CRT-SAME: crtfastmath.o
+! CHECK-CRT-SAME: crtendS.o

flang/test/Driver/frontend-forwarding.f90

 ! Test that flang-new forwards Flang frontend
 ! options to flang-new -fc1 as expected.
 
 ! RUN: %flang -fsyntax-only -### %s -o %t 2>&1 \
 ! RUN:     -finput-charset=utf-8 \
 ! RUN:     -fdefault-double-8 \
 ! RUN:     -fdefault-integer-8 \
 ! RUN:     -fdefault-real-8 \
 ! RUN:     -flarge-sizes \
 ! RUN:     -fconvert=little-endian \
 ! RUN:     -ffp-contract=fast \
-! RUN:     -fno-honor-infinities \
 ! RUN:     -fno-honor-nans \
 ! RUN:     -fapprox-func \
 ! RUN:     -fno-signed-zeros \
 ! RUN:     -fassociative-math \
 ! RUN:     -freciprocal-math \
 ! RUN:     -fpass-plugin=Bye%pluginext \
 ! RUN:     -mllvm -print-before-all\
 ! RUN:     -P \
 ! RUN:   | FileCheck %s
 
 ! CHECK: "-P"
 ! CHECK: "-finput-charset=utf-8"
 ! CHECK: "-fdefault-double-8"
 ! CHECK: "-fdefault-integer-8"
 ! CHECK: "-fdefault-real-8"
 ! CHECK: "-flarge-sizes"
 ! CHECK: "-ffp-contract=fast"
-! CHECK: "-menable-no-infs"
 ! CHECK: "-menable-no-nans"
 ! CHECK: "-fapprox-func"
 ! CHECK: "-fno-signed-zeros"
 ! CHECK: "-mreassociate"
 ! CHECK: "-freciprocal-math"
 ! CHECK: "-fconvert=little-endian"
 ! CHECK: "-fpass-plugin=Bye
 ! CHECK: "-mllvm" "-print-before-all"

flang/test/Lower/fast-math-arithmetic.f90

 ! RUN: %flang_fc1 -emit-fir -ffp-contract=fast %s -o - 2>&1 | FileCheck --check-prefixes=CONTRACT,ALL %s
 ! RUN: %flang_fc1 -emit-fir -menable-no-infs -ffp-contract=off %s -o - 2>&1 | FileCheck --check-prefixes=NINF,ALL %s
 ! RUN: %flang_fc1 -emit-fir -menable-no-nans -ffp-contract=off %s -o - 2>&1 | FileCheck --check-prefixes=NNAN,ALL %s
 ! RUN: %flang_fc1 -emit-fir -fapprox-func -ffp-contract=off %s -o - 2>&1 | FileCheck --check-prefixes=AFN,ALL %s
 ! RUN: %flang_fc1 -emit-fir -fno-signed-zeros -ffp-contract=off %s -o - 2>&1 | FileCheck --check-prefixes=NSZ,ALL %s
 ! RUN: %flang_fc1 -emit-fir -mreassociate -ffp-contract=off %s -o - 2>&1 | FileCheck --check-prefixes=REASSOC,ALL %s
 ! RUN: %flang_fc1 -emit-fir -freciprocal-math -ffp-contract=off %s -o - 2>&1 | FileCheck --check-prefixes=ARCP,ALL %s
 ! RUN: %flang_fc1 -emit-fir -ffp-contract=fast -menable-no-infs -menable-no-nans -fapprox-func -fno-signed-zeros -mreassociate -freciprocal-math %s -o - 2>&1 | FileCheck --check-prefixes=FAST,ALL %s
+! RUN: %flang_fc1 -emit-fir -ffast-math %s -o - 2>&1 | FileCheck --check-prefixes=FAST,ALL %s
 
 ! ALL-LABEL: func.func @_QPtest
 subroutine test(x)
   real x
 ! CONTRACT: arith.mulf{{.*}}, {{.*}} fastmath<[[ATTRS:contract]]> : f32
 ! NINF: arith.mulf{{.*}}, {{.*}} fastmath<[[ATTRS:ninf]]> : f32
 ! NNAN: arith.mulf{{.*}}, {{.*}} fastmath<[[ATTRS:nnan]]> : f32
 ! AFN: arith.mulf{{.*}}, {{.*}} fastmath<[[ATTRS:afn]]> : f32