Thanks for the update. On Thu, Jun 29, 2023, 10:08 PM Jason Lowe-Power <[email protected]> wrote:
> Full system mode is fully supported for RISC-V. There should not be any > problems with running pthreads on RISC-V. > > (Side note: There could be a small bug in the RISC-V atomic > instructions... we're looking into it, but haven't been able to track it > down.) > > Cheers, > Jason > > On Thu, Jun 29, 2023 at 4:05 AM Srikanta Chaitanya via gem5-users < > [email protected]> wrote: > >> Thanks for the reply. >> >> On Thu, Jun 29, 2023, 2:11 PM Víctor Soria <[email protected]> wrote: >> >>> Hi Srikanta, >>> >>> In the past, I've been able to run pthreads mutex with ruby using Arm >>> and FS mode. By default, pthreads use syscalls to the OS to sleep or >>> wake up threads. As far as I know, FS does not work with RISCV. >>> Therefore, you could try to select the pthread mutex type to spinlock to >>> avoid having sleeps and wakes up from the OS. Notice that if you set >>> spinlock, you should never instantiate more threads than cores >>> (oversuscription). Otherwise you can have temporal deadlocks and >>> starvation. >>> >>> Regards, >>> >>> Víctor. >>> >>> On 29/6/23 9:03, Srikanta Chaitanya via gem5-users wrote: >>> > Hi , >>> > I am running pthreads matmul program for riscv in se mode , but with >>> > lock and unlock instructions I am getting read error in Ruby .. >>> > Do i have to compile with mthreads .. >>> > >>> > Thanks >>> > Chaitanya >>> > >>> > _______________________________________________ >>> > gem5-users mailing list -- [email protected] >>> > To unsubscribe send an email to [email protected] >>> >> _______________________________________________ >> gem5-users mailing list -- [email protected] >> To unsubscribe send an email to [email protected] >> >
_______________________________________________ gem5-users mailing list -- [email protected] To unsubscribe send an email to [email protected]
