On 05/06/13 21:27, Matthew Gretton-Dann wrote:

Kugan,

I don't have the source code to hand but how are the sin()/cos()->sincos() optimizations handled?

Thanks Matt. There is a tree level pass to combine sin()/cos() into sincos(). Commit that added this is: http://gcc.gnu.org/viewcvs/gcc?view=revision&revision=121052. We can try doing same thing similar here.

Is there anyway we can know in the tree level that the target does not define integer divide?

Thanks, Kugan

Thanks,

Matt

On 5 June 2013 11:44, Kugan kugan.vivekanandarajah@linaro.org wrote:

...

Hi,

I am looking at best approach for http://gcc.gnu.org/bugzilla/show_bug.cgi?id=43721 - Failure to optimise (a/b) and (a%b) into single __aeabi_idivmod call in ARM architecture

In sumary, the following c code results in __aeabi_idivmod() call and one __aeabi_idiv() call even though the former already calculates the quotient. int q = a / b; int r = a % b; return q + r;

My question is what would be the best way to handle it. As I see there are few options with some issues.

1. Handling in gimple level, try to reduce the operations to equivalent of

this. We should do this for the targets without integer divide. {q, r} = a % b; Gimple assign stmts have only one lhs operation (?). Therefore, lhs has to be made 64bit to signify return values of R0 and R1 returned together. I am not too sure of any implications on other architectures here.

2. Handling in expand_divmod. Here, when we see a div or mod operation, we

will have to do a linear search to see if there is a valid equivalent operation to combine. If we find one, we can generate __aeabi_idivmod() and cache the result for the equivalent operation. As I see, this can get messy and might not be acceptable.

3. An RTL pass to process and combine these library calls. Possibly using

cse. I am still looking at this.

4. Ramana tried a prototype to do the same using target pattens. He has

ruled this out. (if you want more info, please refer to at https://code.launchpad.net/~ramana/gcc-linaro/divmodsi4-experiments)

Any suggestion for best way to handle this?

Thanks, Kugan

On 06/06/13 22:03, Mans Rullgard wrote:

On 6 June 2013 12:09, Kugan kugan.vivekanandarajah@linaro.org wrote:

...

On 05/06/13 21:27, Matthew Gretton-Dann wrote:

...

Kugan,

I don't have the source code to hand but how are the sin()/cos()->sincos() optimizations handled?

Thanks Matt. There is a tree level pass to combine sin()/cos() into sincos(). Commit that added this is: http://gcc.gnu.org/viewcvs/gcc?view=revision&revision=121052. We can try doing same thing similar here.

Is there anyway we can know in the tree level that the target does not define integer divide?

Some targets, e.g. MIPS, have a combined div/mod instruction. Those could benefit from this as well, unless they already achieve that optimisation differently.

Thanks Mans. I will have a look at MIPS port.

Availability of sincos in the target is specified with the define TARGET_HAS_SINCOS and it is being used in tree level optimizations. We can do something similar (?) to get the information about availability of ops/run time library call to consider optimizing.

Thanks, Kugan