Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]
Groups > comp.programming.threads > #3704
| From | Ramine <toto@toto.net> |
|---|---|
| Newsgroups | comp.programming.threads |
| Subject | Re: So stay tuned ! |
| Date | 2017-03-19 15:03 -0400 |
| Organization | A noiseless patient Spider |
| Message-ID | <oamkgh$e7j$3@dont-email.me> (permalink) |
| References | <oamk39$chd$3@dont-email.me> |
On 3/19/2017 2:56 PM, Ramine wrote: > Hello...... > > You don't understand, i was aware that memory ordering is not the only > issue and that you have to use fences and memory barriers but > what i am explaining to you is that my implementation of my algorithms > like my MLock and AMLock etc. already uses the correct fences and memory > barriers to ensure sequential consistency in the hardware side, but what > i need is to switch the optimization off to avoid memory ordering of the > Delphi and FreePascal compilers. > I will switch the optimization off just locally on my MLock and AMLock and some others, that's easy to do to avoid memory reordering of Delphi and FreePascal.. > That's easy to do, i have thought that Delphi and FreePascal doesn't > reorder loads and stores, but this was a mistake, but this mistake is > easy to correct, but mistakes in software programming can happen, but i > am confident that as soon as i correct this mistake, you will be happy > with my C++ Libraries. > > > So stay tuned ! > > > Thank you, > Amine Moulay Ramdne.
Back to comp.programming.threads | Previous | Next — Previous in thread | Find similar | Unroll thread
So stay tuned ! Ramine <toto@toto.net> - 2017-03-19 14:56 -0400 Re: So stay tuned ! Ramine <toto@toto.net> - 2017-03-19 15:03 -0400
csiph-web