[Proposal] Introducing the concept of invalid costs to the IR cost model

Hi,

I’d like to propose a change to our cost interfaces so that instead of returning

an unsigned value from functions like getInstructionCost, getUserCost, etc., we

instead return a wrapper class that encodes an integer cost along with extra

state. The extra state can be used to express:

  1. A cost as infinitely expensive in order to prevent certain optimisations

taking place. For example, there are already examples in LLVM where the cost is

set extremely high, but not so high that it would cause overflow. This might be to

prevent vectorisation in cases where we would have to scalarize the operation,

which is particularly relevant for scalable vectors, where scalarisation is

not [yet] available. There isn’t currently a standard value for something that

constitutes very expensive and we can replace all the magic numbers with a

single invalid state.

  1. A cost as unknown, where the user is simply unable to determine an accurate

cost for an operation.

This new wrapper class would work almost seamlessly with existing code as it

would contain the full set of operators required for arithmetic and comparisons.

This is in addition to the ability to create invalid costs and query the validity

of an existing cost. Once a cost becomes invalid or unknown it will remain in that

state regardless of any further arithmetic performed.

Kind Regards,

David Sherwood.

This sounds like a great step to me David,

-Chris

You’re being quite vague about what the extra state is. Is this because you haven’t yet settled on what it should be, or do you actually intend to let it be a template parameter:

template

struct Cost {

unsigned Val;

T State;

}

Thanks,

Christopher Tetreault

Hi Chris,

Thanks for the reply. My email was really to just ask if people were ok with the idea

in principle and I kind of deliberately avoided getting into too many details at this

point. If people are happy with the idea in principle I can push a patch up soon that

shows what such a class could like so that people can comment on it?

Regards,

David Sherwood.

If people are happy with the idea in principle I can push a patch up soon that

shows what such a class could like so that people can comment on it?

Yeah, or if you got something already but needs a some more work, just upload it and tag it as a work-in-progress patch so people can have a sneak peek what’s coming.

Cheers,

Sjoerd.

Hi David,

This would be a very useful upgrade to the cost model.

One thing I want to add is that we need to be mindful of the cases where the cost is proportional (or inversely proportional) to the VF, for instance in the LoopVectorizationCostModel::selectVectorizationFactor(ElementCount MaxVF), there is a point where expected cost is divided by the VF. I believe there are other places where the instruction cost is dependent on the actual number of elements in the vector. While this is not a problem for fixed vectors, for scalable vectors we need to account for the vscale component of the VF. I guess using the polynomial type for the base cost type might work.

Also, for scalable vectors, when making the cost based decision of whether to vectorize or not, we will need to consider the special case of comparing the scalable vectorization cost with the scalar loop cost, which is considered to be a fixed vector cost of VF=1. A simple solution might be to always assume that for scalable vectors, vectorization is always beneficial. Another option might be to assume that the vscale value is always at least 2, and compare the cost with the scalar loop based on that.

Thanks and Regards,

Vineet

WARNING / LEGAL TEXT: This message is intended only for the use of the individual or entity to which it is addressed and may contain information which is privileged, confidential, proprietary, or exempt from disclosure under applicable law. If you are not the intended recipient or the person responsible for delivering the message to the intended recipient, you are strictly prohibited from disclosing, distributing, copying, or in any way using this message. If you have received this communication in error, please notify the sender and destroy and delete any copies you may have received.

I don't have a lot at stake here, just wanted to point out:

This new wrapper class would work almost seamlessly with existing code as it
would contain the full set of operators required for arithmetic and comparisons.
This is in addition to the ability to create invalid costs and query the validity
of an existing cost. Once a cost becomes invalid or unknown it will remain in that
state regardless of any further arithmetic performed.

Enough thought needs to go into the corner cases. For example, what is
0 * infinite? What is unknown < infinite? What is unknown == unknown?
And so on. It reminds me of the mess that is floating point NaN.

Cheers,
Nicolai