<html>
<head>
<meta content="text/html; charset=utf-8" http-equiv="Content-Type">
</head>
<body bgcolor="#FFFFFF" text="#000000">
<p>TTM limits system memory allocations to 1/2 system memory for
regular processes, or 3/4 for privileged processes.</p>
<p><br>
</p>
<p>For compute we opted to make the GART bigger than system memory
to allow for fragmentation. However, we argued that you don't
really need GART for most buffers anyways, since they are never
accessed in VMID 0. They need to be mapped in some VM page table,
but never in the GART table. So I think the long-term goal should
be to remove the need to allocate GART addresses for system memory
BOs. Then you could get away with a quite small GART size.<br>
</p>
<p><br>
</p>
<p>Regards,</p>
<p> Felix<br>
</p>
<br>
<div class="moz-cite-prefix">On 16-07-21 08:04 AM, Christian König
wrote:<br>
</div>
<blockquote
cite="mid:3839ef42-6275-99d8-9aee-54143a971f0b@vodafone.de"
type="cite">
<meta http-equiv="Content-Type" content="text/html; charset=utf-8">
<div class="moz-cite-prefix">
<blockquote type="cite">
<p>we can add people in if necessary here. so the initial
though of this change is with latest big memory GPU
released, the remote memory goes very large, like Polaris10.</p>
<p>According to the windows WDDM, remote memory size is half
of the system memory, I don't know what was discussed
before, Please let me know if there're use case indeed need
that big of remote memory.</p>
</blockquote>
Well that is at least something that makes sense.<br>
<br>
Previous discussions where always settled around making GART
bigger than VRAM and not the other way around.<br>
<br>
Anyway I think a prerequisite to this is actually allowing half
of system memory to be allocated for GART and I'm not sure if
that is currently the case. TTM has a limitation for that as
well, but I off hand don't remember how high much it was.<br>
<br>
I think we should use the same value which TTM uses for the
upper limit here.<br>
<br>
Regards,<br>
Christian.<br>
<br>
Am 21.07.2016 um 13:56 schrieb Wang, Qingqing:<br>
</div>
<blockquote
cite="mid:SN1PR12MB0511DD6F60EAFDB326320B46EC090@SN1PR12MB0511.namprd12.prod.outlook.com"
type="cite">
<style type="text/css" style="display:none;"><!-- P {margin-top:0;margin-bottom:0;} --></style>
<div id="divtagdefaultwrapper"
style="font-size:12pt;color:#000000;background-color:#FFFFFF;font-family:Calibri,Arial,Helvetica,sans-serif;">
<p>alright, I didn't realize you're not in that thread. </p>
<p>we can add people in if necessary here. so the initial
though of this change is with latest big memory GPU
released, the remote memory goes very large, like Polaris10.</p>
<p>According to the windows WDDM, remote memory size is half
of the system memory, I don't know what was discussed
before, Please let me know if there're use case indeed need
that big of remote memory.</p>
</div>
<hr style="display:inline-block;width:98%" tabindex="-1">
<div id="divRplyFwdMsg" dir="ltr"><font style="font-size:11pt"
color="#000000" face="Calibri, sans-serif"><b>发件人:</b>
Christian König <a moz-do-not-send="true"
class="moz-txt-link-rfc2396E"
href="mailto:deathsimple@vodafone.de"><deathsimple@vodafone.de></a><br>
<b>发送时间:</b> 2016年7月21日 19:48:17<br>
<b>收件人:</b> Wang, Qingqing; <a moz-do-not-send="true"
class="moz-txt-link-abbreviated"
href="mailto:amd-gfx@lists.freedesktop.org">amd-gfx@lists.freedesktop.org</a><br>
<b>主题:</b> Re: 答复: 答复: [PATCH] drm/amdgpu: adjust gtt memory
size</font>
<div> </div>
</div>
<div>
<div class="moz-cite-prefix">Am 21.07.2016 um 13:15 schrieb
Wang, Qingqing:<br>
</div>
<blockquote
cite="mid:SN1PR12MB051120166EEFC6A65C7BB795EC090@SN1PR12MB0511.namprd12.prod.outlook.com"
type="cite">
<style type="text/css" style="display:none;"><!-- P {margin-top:0;margin-bottom:0;} --></style>
<div id="divtagdefaultwrapper"
style="font-size:12pt;color:#000000;background-color:#FFFFFF;font-family:Calibri,Arial,Helvetica,sans-serif;">
<div id="divtagdefaultwrapper"
style="font-size:12pt;color:#000000;background-color:#FFFFFF;font-family:Calibri,Arial,Helvetica,sans-serif;">
<p><span style="color: rgb(33, 33, 33); font-size:
15px;">Actually that discussion was held long ago
internally but we never got a conclusion on this
IIRC. So we should probably continue the discussion
on this thread now.</span><br>
</p>
<p><font color="#212121"><span style="font-size: 15px;">-there
is a thread for this topic yesterday, some people
are not in amd-gfx mail list, we'd
better discuss in the internal thread.</span></font></p>
</div>
</div>
</blockquote>
No, exactly for this reason we have the public mailing list. I
for example wasn't part of the internal thread either.<br>
<br>
Christian.<br>
<br>
<blockquote
cite="mid:SN1PR12MB051120166EEFC6A65C7BB795EC090@SN1PR12MB0511.namprd12.prod.outlook.com"
type="cite">
<div id="divtagdefaultwrapper"
style="font-size:12pt;color:#000000;background-color:#FFFFFF;font-family:Calibri,Arial,Helvetica,sans-serif;">
<hr style="display:inline-block;width:98%" tabindex="-1">
<div id="divRplyFwdMsg" dir="ltr"><font
style="font-size:11pt" color="#000000" face="Calibri,
sans-serif"><b>发件人:</b> Christian König <a
moz-do-not-send="true" class="moz-txt-link-rfc2396E"
href="mailto:deathsimple@vodafone.de"><deathsimple@vodafone.de></a><br>
<b>发送时间:</b> 2016年7月21日 17:48:41<br>
<b>收件人:</b> Wang, Qingqing; <a moz-do-not-send="true"
class="moz-txt-link-abbreviated"
href="mailto:amd-gfx@lists.freedesktop.org">
amd-gfx@lists.freedesktop.org</a><br>
<b>主题:</b> Re: 答复: [PATCH] drm/amdgpu: adjust gtt
memory size</font>
<div> </div>
</div>
<div>
<div class="moz-cite-prefix">Am 21.07.2016 um 11:16
schrieb Wang, Qingqing:<br>
</div>
<blockquote
cite="mid:SN1PR12MB0511612A271156A2EFC7B449EC090@SN1PR12MB0511.namprd12.prod.outlook.com"
type="cite">
<meta name="Generator" content="Microsoft Exchange
Server">
<!-- converted from text -->
<meta content="text/html; charset=UTF-8">
<div dir="ltr">
<div id="x_divtagdefaultwrapper"
style="font-size:12pt; color:#000000;
background-color:#FFFFFF;
font-family:Calibri,Arial,Helvetica,sans-serif">
<p><font size="2"><span style="font-size:10pt">This
doesn't work on 32bit systems since totalram
is a long (IIRC) and <br>
so this will certainly overflow.</span></font></p>
<p>-- can you be more specific, how could the
overflow happen?</p>
</div>
</div>
</blockquote>
<br>
Both values are 32bit on a 32bit system, so as soon as
you have more than 4GB installed this will overflow.<br>
<br>
You need to cast the long to a 64bit value to avoid
that.<br>
<br>
<blockquote
cite="mid:SN1PR12MB0511612A271156A2EFC7B449EC090@SN1PR12MB0511.namprd12.prod.outlook.com"
type="cite">
<div dir="ltr">
<div id="x_divtagdefaultwrapper"
style="font-size:12pt; color:#000000;
background-color:#FFFFFF;
font-family:Calibri,Arial,Helvetica,sans-serif">
<p><br>
</p>
<p><font size="2"><span style="font-size:10pt">Additional
if I remember correctly we didn't ended the
discussion on <br>
what to do here with a conclusion.</span></font></p>
<p>-- ok, since it happens in anther mail list, we
should stop talking about it here, you can send
out your ideas there.<br>
</p>
</div>
</div>
</blockquote>
<br>
Actually that discussion was held long ago internally
but we never got a conclusion on this IIRC. So we should
probably continue the discussion on this thread now.<br>
<br>
Regards,<br>
Christian.<br>
<br>
<blockquote
cite="mid:SN1PR12MB0511612A271156A2EFC7B449EC090@SN1PR12MB0511.namprd12.prod.outlook.com"
type="cite">
<div dir="ltr">
<div id="x_divtagdefaultwrapper"
style="font-size:12pt; color:#000000;
background-color:#FFFFFF;
font-family:Calibri,Arial,Helvetica,sans-serif"> </div>
<hr tabindex="-1" style="display:inline-block;
width:98%">
<div id="x_divRplyFwdMsg" dir="ltr"><font
style="font-size:11pt" color="#000000"
face="Calibri, sans-serif"><b>发件人:</b> amd-gfx <a
moz-do-not-send="true"
class="moz-txt-link-rfc2396E"
href="mailto:amd-gfx-bounces@lists.freedesktop.org">
<amd-gfx-bounces@lists.freedesktop.org></a>
代表 Christian König <a moz-do-not-send="true"
class="moz-txt-link-rfc2396E"
href="mailto:deathsimple@vodafone.de">
<deathsimple@vodafone.de></a><br>
<b>发送时间:</b> 2016年7月21日 16:53:05<br>
<b>收件人:</b> Wang, Qingqing; <a
moz-do-not-send="true"
class="moz-txt-link-abbreviated"
href="mailto:amd-gfx@lists.freedesktop.org">
amd-gfx@lists.freedesktop.org</a><br>
<b>主题:</b> Re: [PATCH] drm/amdgpu: adjust gtt
memory size</font>
<div> </div>
</div>
</div>
<font size="2"><span style="font-size:10pt;">
<div class="PlainText">Am 21.07.2016 um 09:46
schrieb Ken Wang:<br>
> Change-Id:
If00d5b97ba9e30f9b7f68fdfc134a0f8b3ad2add<br>
> Signed-off-by: Ken Wang <a
moz-do-not-send="true"
class="moz-txt-link-rfc2396E"
href="mailto:Qingqing.Wang@amd.com">
<Qingqing.Wang@amd.com></a><br>
> ---<br>
> drivers/gpu/drm/amd/amdgpu/gmc_v7_0.c | 7
++++++-<br>
> drivers/gpu/drm/amd/amdgpu/gmc_v8_0.c | 7
++++++-<br>
> 2 files changed, 12 insertions(+), 2
deletions(-)<br>
><br>
> diff --git
a/drivers/gpu/drm/amd/amdgpu/gmc_v7_0.c
b/drivers/gpu/drm/amd/amdgpu/gmc_v7_0.c<br>
> index a3b6048..0b095d5 100644<br>
> --- a/drivers/gpu/drm/amd/amdgpu/gmc_v7_0.c<br>
> +++ b/drivers/gpu/drm/amd/amdgpu/gmc_v7_0.c<br>
> @@ -336,6 +336,11 @@ static int
gmc_v7_0_mc_init(struct amdgpu_device *adev)<br>
> {<br>
> u32 tmp;<br>
> int chansize, numchan;<br>
> + u64 sysmem_size;<br>
> + struct sysinfo si;<br>
> +<br>
> + si_meminfo(&si);<br>
> + sysmem_size = si.totalram *
si.mem_unit;<br>
<br>
This doesn't work on 32bit systems since
totalram is a long (IIRC) and <br>
so this will certainly overflow.<br>
<br>
Additional if I remember correctly we didn't
ended the discussion on <br>
what to do here with a conclusion.<br>
<br>
Regards,<br>
Christian.<br>
<br>
> <br>
> /* Get VRAM informations */<br>
> tmp = RREG32(mmMC_ARB_RAMCFG);<br>
> @@ -392,7 +397,7 @@ static int
gmc_v7_0_mc_init(struct amdgpu_device *adev)<br>
> * size equal to the 1024 or vram,
whichever is larger.<br>
> */<br>
> if (amdgpu_gart_size == -1)<br>
> - adev->mc.gtt_size =
max((1024ULL << 20),
adev->mc.mc_vram_size);<br>
> + adev->mc.gtt_size =
max((1024ULL << 20),
min(adev->mc.mc_vram_size, sysmem_size/2));<br>
> else<br>
> adev->mc.gtt_size =
(uint64_t)amdgpu_gart_size << 20;<br>
> <br>
> diff --git
a/drivers/gpu/drm/amd/amdgpu/gmc_v8_0.c
b/drivers/gpu/drm/amd/amdgpu/gmc_v8_0.c<br>
> index 02b6872..790bf7a 100644<br>
> --- a/drivers/gpu/drm/amd/amdgpu/gmc_v8_0.c<br>
> +++ b/drivers/gpu/drm/amd/amdgpu/gmc_v8_0.c<br>
> @@ -413,6 +413,11 @@ static int
gmc_v8_0_mc_init(struct amdgpu_device *adev)<br>
> {<br>
> u32 tmp;<br>
> int chansize, numchan;<br>
> + u64 sysmem_size;<br>
> + struct sysinfo si;<br>
> +<br>
> + si_meminfo(&si);<br>
> + sysmem_size = si.totalram *
si.mem_unit;<br>
> <br>
> /* Get VRAM informations */<br>
> tmp = RREG32(mmMC_ARB_RAMCFG);<br>
> @@ -469,7 +474,7 @@ static int
gmc_v8_0_mc_init(struct amdgpu_device *adev)<br>
> * size equal to the 1024 or vram,
whichever is larger.<br>
> */<br>
> if (amdgpu_gart_size == -1)<br>
> - adev->mc.gtt_size =
max((1024ULL << 20),
adev->mc.mc_vram_size);<br>
> + adev->mc.gtt_size =
max((1024ULL << 20),
min(adev->mc.mc_vram_size, sysmem_size/2));<br>
> else<br>
> adev->mc.gtt_size =
(uint64_t)amdgpu_gart_size << 20;<br>
> <br>
<br>
<br>
_______________________________________________<br>
amd-gfx mailing list<br>
<a moz-do-not-send="true"
class="moz-txt-link-abbreviated"
href="mailto:amd-gfx@lists.freedesktop.org">amd-gfx@lists.freedesktop.org</a><br>
<a moz-do-not-send="true"
href="https://lists.freedesktop.org/mailman/listinfo/amd-gfx">https://lists.freedesktop.org/mailman/listinfo/amd-gfx</a><br>
</div>
</span></font></blockquote>
<p><br>
</p>
</div>
</div>
</blockquote>
<p><br>
</p>
</div>
</blockquote>
<p><br>
</p>
<br>
<fieldset class="mimeAttachmentHeader"></fieldset>
<br>
<pre wrap="">_______________________________________________
amd-gfx mailing list
<a class="moz-txt-link-abbreviated" href="mailto:amd-gfx@lists.freedesktop.org">amd-gfx@lists.freedesktop.org</a>
<a class="moz-txt-link-freetext" href="https://lists.freedesktop.org/mailman/listinfo/amd-gfx">https://lists.freedesktop.org/mailman/listinfo/amd-gfx</a>
</pre>
</blockquote>
<br>
</body>
</html>