IBM Looks To Squeeze Support For Future Power "Dense Math" CPUs Into GCC 13

Back in November IBM engineers sent out compiler patches for "future" CPUs and notably added new dense math instructions. At the time they noted the new target "may or may not be present" in future IBM Power processors but now they are hoping to get these "future" patches squeeze into the GCC 13 compiler.

While they are cautious with their wording, this IBM "future" CPU support is likely what may end up being launched as IBM Power11 processors. IBM did a similar "future" ploy with their Power10 compiler support.

Atop Power10, the primary aim with these "future" patches for the GNU Compiler Collection was new dense math instructions.The future CPUs add new dense math registers and other changes to make on the compiler side for properly targeting those "future" processors.

Since November there haven't been any major new compiler patch series from IBM while the GCC 13 compiler release is coming up in March~April. This week there were "pings" over the November patches with hopes of getting them into this next compiler version.

Ping patch. We really would like the patches to enable the possible future MMA+ instructions into GCC 13.

So while the future support was talked down with the usual legal "may or may not be in actual PowerPCs
that will be delivered in the future", it seems pretty much a given. And with IBM wanting to get these patches into GCC 13 rather than waiting until GCC 14 next year gives hopes the new processors could be announced this year. In any event when the new processors are formally released, IBM goes back and updates the "-mcpu=future" target to properly reflect the branding of the released processors.