Thank you! Wonderful. This is probably the BCD Assist that is mentioned in
https://www.websphereusergroup.co.uk/wug/presentations/40/Whats_new_In_IBM_Java_8.pdf
.
René.
> On 23 Mar 2023, at 03:29, Andrew Rowley wrote:
>
> On 23/03/2023 11:39 am, David Crayford wrote:
>> FYI, the OpenJ9 JIT ca
On 23/3/23 10:29, Andrew Rowley wrote:
On 23/03/2023 11:39 am, David Crayford wrote:
FYI, the OpenJ9 JIT can implement JIT intrinsics to generate native
decimal instructions when running on z/OS. IBM have open source
OpenJ9 and OMR so you can snoop around the code
https://github.com/eclipse-op
On 23/03/2023 11:39 am, David Crayford wrote:
FYI, the OpenJ9 JIT can implement JIT intrinsics to generate native
decimal instructions when running on z/OS. IBM have open source OpenJ9
and OMR so you can snoop around the code
https://github.com/eclipse-openj9/openj9/blob/master/jcl/src/openj9.d
On 23/3/23 01:03, René Jansen wrote:
That's another interesting take; but I have to be sure that it is used before I
declare the winner.
I stand corrected. I just read the hardware specs. DFP was originally
implemented in millicode (vertical microcode) on the z9 but has been on
the hardware
On 23/03/2023 3:42 am, René Jansen wrote:
Does anybody know how to ask the J9 (Java 8) on z/OS how to show me what it
does when the JIT decides native code would be best?
I think it is difficult to pin down the native code because the JIT is
not a one-time, fixed output compiler.
Java can mon
On Wed, 22 Mar 2023 at 14:46, René Jansen wrote:
[...]
> So I am setting up a few benchmarks, and I will report back. But the first
> question is, how does it work?
> Java does not have a normal compiler backend, where we can tell it to compile
> for a certain ISA
> (or default to the one it is
I benchmarked DFP extensively when it was first released (decades ago
now). For functional equivalency (same precision in result) it is 100x
faster than the best non-DFP optimized implementation. For instruction
equivalency, it is slightly faster than HFP and BFP. There was no
condition when DFP
When we compare the DFP instructions to the 'old' floating point instructions,
we can see that these were storage-to-storage type instructions, while the new
set is register-to-register, which can make an enormous difference with fast
instruction caches and -pipelines -- this is why my C code is
Mainframe Discussion List [IBM-MAIN@LISTSERV.UA.EDU] on behalf of
René Jansen [rene.vincent.jan...@gmail.com]
Sent: Wednesday, March 22, 2023 12:42 PM
To: IBM-MAIN@LISTSERV.UA.EDU
Subject: watching JIT generated machine instructions on z/OS
Without reading any documentation (sorry!), the issue at hand is
The difference may be down at the noise level - unless you are doing
millions of these a second.
For example if you had a *load register,address* in a tight loop- the
second time may be much (100?) faster because the conversion of virtual
address to real page address will already be done, and the d
That's another interesting take; but I have to be sure that it is used before I
declare the winner.
> On 22 Mar 2023, at 18:00, David Crayford wrote:
>
> I can't answer you original question but I doubt if DFP is really that much
> faster. I would imagine it's implemented in millicode and not
I can't answer you original question but I doubt if DFP is really that
much faster. I would imagine it's implemented in millicode and not
silicone so is a software implementation at heart. I would be surprised
if it beats BigDecimal on a PC but I could be wrong.
On 23/3/23 00:42, René Jansen w
Without reading any documentation (sorry!), the issue at hand is this. I want
to show the performance gains of using DFP (Decimal Floating Point) for the
typical financial application, after I noticed at some other client their
bought packages seldom were compiled using the right compiler option
13 matches
Mail list logo