By: zArchJon (jdbradbu.delete@this.us.ibm.com), August 24, 2022 10:12 am
Room: Moderated Discussions
rwessel (rwessel.delete@this.yahoo.com) on August 20, 2022 6:50 pm wrote:
> HTM on about the only platform on which it's been reasonably
> successfully implemented, namely Z, is going away.
>
> On page 93 of:
>
> https://www.redbooks.ibm.com/redpieces/pdfs/sg248951.pdf
>
> "Removal of support of the transactional execution and constrained transactional
> execution facility: In a future IBM Z hardware system family, the transactional execution
> and constrained transactional execution facility will no longer be supported. Users of the
> facility on current servers should always check the facility indications before use."
>
> So introduced on the EC12s, and still on z16, but not long for this world.
>
> IBM also removed it from Power10 a couple of years ago.
It is true that HTM has had considerably less positive impact than we had hoped for a decade ago, and it’s also true that it is a significant effort to keep it working correctly when changing processor architecture for other reasons. For that reason, we have announced that at some future point we will stop supporting HTM. To be clear, the z16 generation is still fully supporting HTM, and we are working internally at IBM and with other software vendors and clients to ensure we have adequate technologies (e.g. new instructions supporting complex locking scenarios) in place for when we remove HTM.
IBM has a history of complex atomic operations (compare and swap was invented for the S/370). We currently have compare and swap, compare double and swap, and compare and swap and store instructions. For more complex locked updates we have the Perform Locked Operation instruction which helps to solve the problem of being preempted while holding the lock since the lock is an internal machine owned lock.
> HTM on about the only platform on which it's been reasonably
> successfully implemented, namely Z, is going away.
>
> On page 93 of:
>
> https://www.redbooks.ibm.com/redpieces/pdfs/sg248951.pdf
>
> "Removal of support of the transactional execution and constrained transactional
> execution facility: In a future IBM Z hardware system family, the transactional execution
> and constrained transactional execution facility will no longer be supported. Users of the
> facility on current servers should always check the facility indications before use."
>
> So introduced on the EC12s, and still on z16, but not long for this world.
>
> IBM also removed it from Power10 a couple of years ago.
It is true that HTM has had considerably less positive impact than we had hoped for a decade ago, and it’s also true that it is a significant effort to keep it working correctly when changing processor architecture for other reasons. For that reason, we have announced that at some future point we will stop supporting HTM. To be clear, the z16 generation is still fully supporting HTM, and we are working internally at IBM and with other software vendors and clients to ensure we have adequate technologies (e.g. new instructions supporting complex locking scenarios) in place for when we remove HTM.
IBM has a history of complex atomic operations (compare and swap was invented for the S/370). We currently have compare and swap, compare double and swap, and compare and swap and store instructions. For more complex locked updates we have the Perform Locked Operation instruction which helps to solve the problem of being preempted while holding the lock since the lock is an internal machine owned lock.
