On Fri, Apr 12, 2013 at 07:10:13PM +0100, Chris Wilson wrote:
> Haswell introduces a separate frequency domain for the ring (uncore). So
> where we used to increase the CPU (IA) clock with GPU busyness, we now
> need to scale the ring frequency directly instead. As the ring limits
> our memory band
Haswell introduces a separate frequency domain for the ring (uncore). So
where we used to increase the CPU (IA) clock with GPU busyness, we now
need to scale the ring frequency directly instead. As the ring limits
our memory bandwidth, it is vital for performance that when the GPU is
busy, we incre
On Wed, Apr 10, 2013 at 02:41:39PM -0700, Jesse Barnes wrote:
> On Wed, 10 Apr 2013 20:45:11 +0100
> Chris Wilson wrote:
>
> > + ring_freq = (gpu_freq * 5 + 3) / 4;
> > + ring_freq = max(max_ring_freq, ring_freq);
>
> Should this be min(max_ring_freq, ring_fre
On Wed, 10 Apr 2013 20:45:11 +0100
Chris Wilson wrote:
> + ring_freq = (gpu_freq * 5 + 3) / 4;
> + ring_freq = max(max_ring_freq, ring_freq);
Should this be min(max_ring_freq, ring_freq)? Otherwise it seems like
every ia_freq will correspond to the highes
On Wed, Apr 10, 2013 at 08:45:11PM +0100, Chris Wilson wrote:
> Haswell introduces a separate frequency domain for the ring (uncore). So
> where we used to increase the CPU (IA) clock with GPU busyness, we now
> need to scale the ring frequency directly instead. As the ring limits
> our memory band
Haswell introduces a separate frequency domain for the ring (uncore). So
where we used to increase the CPU (IA) clock with GPU busyness, we now
need to scale the ring frequency directly instead. As the ring limits
our memory bandwidth, it is vital for performance that when the GPU is
busy, we incre
Haswell introduces a separate frequency domain for the ring (uncore). So
where we used to increase the CPU (IA) clock with GPU busyness, we now
need to scale the ring frequency directly instead. As the ring limits
our memory bandwidth, it is vital for performance that when the GPU is
busy, we incre
