I too am concerned about Metal/C not keeping up with the hardware generations, especially as it seems that IBM is about to announce another generation.
Has anyone already submitted an “idea” to update Metal/C hardware support? Or in the alternative, to open-source Metal/C so that the community can upgrade it themselves? (Yeah, I know – two chances of that happening: slim and none, but it doesn’t hurt to ask.) Peter From: IBM Mainframe Discussion List <IBM-MAIN@LISTSERV.UA.EDU> On Behalf Of David Crayford Sent: Sunday, March 16, 2025 10:39 AM To: IBM-MAIN@LISTSERV.UA.EDU Subject: Re: Open XL C dramaticallly slower thant z/OS XL C compiler - expected? > On 16 Mar 2025, at 3:02 pm, David Cole <dbc...@colesoft.com> wrote: > > I think the more interesting question is not, whether the compiler is slower. > It's whether the compiled code is slower. > > Is it? That’s an interesting question because the legacy XL C/C++ compiler has been functionally stabilized. There is significant concern that Metal/C has not been updated to support the z16 architecture. In our experiments, the Open XL C/C++ Clang compiler is 26 times slower at compiling the same code compared to Clang running on a zCX container on the same LPAR. The binary is also 236M compared to 123K on Linux. This is very much a work in progress. The most concerning issue I have encountered is the limited runtime. For example, features like extended stdio for VSAM and SVC 99 are not enabled. We do not consider this compiler ready for product development. However, it is excellent for porting Linux applications to z/OS. > Dave Cole, Developer > dbc...@gmail.com <mailto:dbc...@gmail.com> (personal) > dbc...@colesoft.com <mailto:dbc...@colesoft.com> (business) > 540-456-6518 (cell) > > At 3/16/2025 01:33 AM, Allan Staller wrote: >> Classification: Confidential >> >> Region. The Optimizer takes a whole lot of virtual storage >> >> -----Original Message----- >> From: IBM Mainframe Discussion List <IBM-MAIN@LISTSERV.UA.EDU >> <mailto:IBM-MAIN@LISTSERV.UA.EDU>> On Behalf Of Charles Mills >> Sent: Friday, March 14, 2025 1:39 PM >> To: IBM-MAIN@LISTSERV.UA.EDU <mailto:IBM-MAIN@LISTSERV.UA.EDU> >> Subject: Open XL C dramaticallly slower thant z/OS XL C compiler - expected? >> >> I am trying to get started with the new, "open," Clang-based XLC compiler. I >> am stunned at how noticeably slower it is than the legacy C compiler. I >> wasn't looking to benchmark -- the slow compile times just jumped out at me. >> >> Is this to be expected? I wish I had numbers for exactly the same source >> code but there are different userids involved and it's a little complicated. >> Different options for the two compiles because the option specifications are >> incompatible. Both modules are fairly vanilla source code written by the >> same author. >> >> Both of the following are on the same machine, the same virtual machine at >> IBM Dallas. It's a z16 A01. >> >> Compile under the legacy XL C compiler, a C++ compile of a 1886-line source >> module: 4 seconds elapsed, .36 CPU seconds. >> >> Compile under Open XL C compiler, a C++ compile of a 415-line source module: >> 26 seconds elapsed, 1.32 CPU seconds. >> >> Is this what others are seeing? Is this to be expected? IS anyone using the >> new compiler? >> >> Charles -- This message and any attachments are intended only for the use of the addressee and may contain information that is privileged and confidential. If the reader of the message is not the intended recipient or an authorized representative of the intended recipient, you are hereby notified that any dissemination of this communication is strictly prohibited. If you have received this communication in error, please notify us immediately by e-mail and delete the message and any attachments from your system. ---------------------------------------------------------------------- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
