[ofa-general] atomic operations on ppc64

[Dotan Barak]

Rui:

Do you access the memory as "volatile" (becuase the HCA changes its content)?
Can you try do check if RDMA Write/Read behave the same for you?

Dotan

2008/9/22 Ronni Zimmermann <ronniz at mellanox.co.il>:
> Hi,
> We run tests which use atomic operations (both fetch and add and comp and swap) on PPC64 all the time, without experiencing any problem.
>
> Just to make sure I ran few simple tests, which use atomic operations, on our PPC64 machines, both with SLES10 SP1 and with RHAS5.1, and all of them passed.
> I was working with the latest OFED1.4 driver and mlx4 HCA with the latest released FW and with FW 2.3.000 (on the SLES10 SP1 machine).
>
> Given the above information I believe that there's either a problem with your code (although looking at the code you posted I couldn't see anything wrong) or it's an OFED1.2.5 issue, as Dotan suggested.
>
> Ronni.
>
>
> -----Original Message-----
> From: general-bounces at lists.openfabrics.org [mailto:general-bounces at lists.openfabrics.org] On Behalf Of Rui Machado
> Sent: ו 19 ספטמבר 2008 17:39
> To: Dotan Barak
> Cc: general at lists.openfabrics.org
> Subject: Re: [ofa-general] atomic operations on ppc64
>
> 2008/9/19 Dotan Barak <dotanba at gmail.com>:
>>
>>>> It seems that you didn't check the same HCA in each arch.
>>>>
>>>> Can you try to use the mthca device in PPC64 and check the results?
>>>>
>>>> (Anyway, i would have suggest you to upgrade the OFED package that
>>>> you are
>>>> using)
>>>>
>>>
>>> I apologize for my newbieness here but I do not understand what you mean.
>>> The two machines have different devices. What do you mean by making
>>> the ppc64 machine use the mthca when it does only have the mlx4?
>>>
>>> Unfortunately, upgrading might not be an (easy) option :-/
>>>
>>> Thanks for the patience
>>>
>>
>> You have 2 machines, each with another device:
>> x86 has the HCA 25204
>> PPC64 has the HCA 25418
>>
>>
>> The fact that the test didn't pass for you on the PPC64 may be related
>> to this device (i think that OFED 1.2.5 was the first version that
>> supported this device).
>>
>> So, i suggest that you'll put the HCA 25204 on the PPC64 and check if
>> the failure still exists.
>>
>>
>> Dotan
>>
>
> Right :) Now I get it!
>
> I couldn't (yet at least) put the HCA 25204 on the ppc64.
>
> But I tried again with another ppc64 machine.
>
> ibv_devinfo
> hca_id: mthca0
>        fw_ver:                         4.8.200
>        node_guid:                      0005:ad00:001d:cd24
>        sys_image_guid:                 0005:ad00:0100:d050
>        vendor_id:                      0x05ad
>        vendor_part_id:                 25208
>        hw_ver:                         0xA0
>        board_id:                       HCA.HSDC.A0
>        phys_port_cnt:                  2
>                port:   1
>                        state:                  PORT_ACTIVE (4)
>                        max_mtu:                2048 (4)
>                        active_mtu:             2048 (4)
>                        sm_lid:                 142
>                        port_lid:               156
>                        port_lmc:               0x01
>
>                port:   2
>                        state:                  PORT_DOWN (1)
>                        max_mtu:                2048 (4)
>                        active_mtu:             512 (2)
>                        sm_lid:                 0
>                        port_lid:               0
>                        port_lmc:               0x00
>
>
> The result was still the same as before:
> ppc64-ppc64 - the counter doesn't get updated.
> ppc64-x86 - clients hangs at poll
>
> Just as an extra info...
> _______________________________________________
> general mailing list
> general at lists.openfabrics.org
> http://lists.openfabrics.org/cgi-bin/mailman/listinfo/general
>
> To unsubscribe, please visit http://openib.org/mailman/listinfo/openib-general
> _______________________________________________
> general mailing list
> general at lists.openfabrics.org
> http://lists.openfabrics.org/cgi-bin/mailman/listinfo/general
>
> To unsubscribe, please visit http://openib.org/mailman/listinfo/openib-general
>