My *impression*. Two disclaimers: I am not an architecture expert and I don't have time at the moment to do any real research.
I believe that the problem for synchronization with STCK is that the lowest order bits are of such relatively great magnitude that a lot of hardware spinning may be required to delay until a unique value can be obtained. STCKE on the other hand has much "lower" low-order bits and so much less spinning is potentially required. In other words, I think STCKE solves the spin problem, but at the cost of being incompatible -- different and larger output field -- than STCK. STCKF on the other hand is plug-compatible with STCK, assuming that you don't need uniqueness, don't need a monotonically increasing value. Charles -----Original Message----- From: IBM Mainframe Assembler List <[email protected]> On Behalf Of Mario Bezzi Sent: Friday, April 11, 2025 10:38 AM To: [email protected] Subject: Re: Store-Clock-Fast Facility Thank you all for your kind help. @Charles: You make a good point about STCKE, but I was under the impression, confirmed by a quick check that as STCK, it also requires synchronization. "A serialization function is performed before the value of the clock is fetched and again after the value is placed in storage" Thanks! mario
