All of lore.kernel.org
 help / color / mirror / Atom feed
From: "Bernatowicz, Marcin" <marcin.bernatowicz@linux.intel.com>
To: Adam Miszczak <adam.miszczak@linux.intel.com>,
	igt-dev@lists.freedesktop.org
Cc: kamil.konieczny@linux.intel.com, michal.wajdeczko@intel.com,
	pawel.sikora@intel.com
Subject: Re: [PATCH i-g-t v3 3/4] tools/vmtb: Test resources - vGPU profiles
Date: Thu, 5 Dec 2024 10:51:22 +0100	[thread overview]
Message-ID: <37e9e22b-01e2-40ca-a761-b50980532be3@linux.intel.com> (raw)
In-Reply-To: <20241127102202.3903735-4-adam.miszczak@linux.intel.com>



On 11/27/2024 11:22 AM, Adam Miszczak wrote:
> Provide vGPU profiles definition for Flex 170 (ATS-M) platform.
> Definitions are provided in a JSON format, parsed by the VMTB
> and applied as a test setup (fixture).
> 
> Signed-off-by: Adam Miszczak <adam.miszczak@linux.intel.com>
> ---
>   .../resources/vgpu_profiles/Flex170.json      | 113 ++++++++++++++++++
>   1 file changed, 113 insertions(+)
>   create mode 100644 tools/vmtb/vmm_flows/resources/vgpu_profiles/Flex170.json
> 
> diff --git a/tools/vmtb/vmm_flows/resources/vgpu_profiles/Flex170.json b/tools/vmtb/vmm_flows/resources/vgpu_profiles/Flex170.json
> new file mode 100644
> index 000000000..ff1fa7e20
> --- /dev/null
> +++ b/tools/vmtb/vmm_flows/resources/vgpu_profiles/Flex170.json
> @@ -0,0 +1,113 @@
> +{
> +    "version": "1.1",
> +    "PFResources": {
> +        "Default": "MinimumPFResources",
> +        "Profile": {
> +            "MinimumPFResources": {
> +                "LocalMemoryEccOn": 402653184,
> +                "LocalMemoryEccOff": 402653184,
> +                "Contexts": 1024,
> +                "Doorbells": 16,
> +                "GGTTSize": 268435456
> +            }
> +        }
> +    },
> +    "vGPUResources": {
> +        "Default": null,
> +        "Profile": {
> +            "Flex170_16": {
> +                "VFCount": 1,
> +                "LocalMemoryEccOff": 16777216000,
> +                "LocalMemoryEccOn": 2147483648,
> +                "Contexts": 1024,
> +                "Doorbells": 240,
> +                "GGTTSize": 4026531840
> +            },
> +            "Flex170_8": {
> +                "VFCount": 2,
> +                "LocalMemoryEccOff": 8388608000,
> +                "LocalMemoryEccOn": 2147483648,
> +                "Contexts": 1024,
> +                "Doorbells": 120,
> +                "GGTTSize": 2013265920
> +            },
> +            "Flex170_4": {
> +                "VFCount": 4,
> +                "LocalMemoryEccOff": 4194304000,
> +                "LocalMemoryEccOn": 2147483648,
> +                "Contexts": 1024,
> +                "Doorbells": 60,
> +                "GGTTSize": 1006632960
> +            },
> +            "Flex170_2": {
> +                "VFCount": 8,
> +                "LocalMemoryEccOff": 2097152000,
> +                "LocalMemoryEccOn": 1073741824,
> +                "Contexts": 1024,
> +                "Doorbells": 30,
> +                "GGTTSize": 503316480
> +            },
> +            "Flex170_1": {
> +                "VFCount": 16,
> +                "LocalMemoryEccOff": 1048576000,
> +                "LocalMemoryEccOn": 536870912,
> +                "Contexts": 1024,
> +                "Doorbells": 15,
> +                "GGTTSize": 251658240
> +            }
> +        }
> +    },
> +    "vGPUScheduler": {
> +        "Default": "Flexible_30fps_GPUTimeSlicing",
> +        "Profile": {
> +            "Flexible_30fps_GPUTimeSlicing": {
> +                "GPUTimeSlicing": {
> +                    "ScheduleIfIdle": false,
> +                    "PFExecutionQuantum": 20,
> +                    "PFPreemptionTimeout": 20000,
> +                    "VFAttributes": {
> +                        "VFExecutionQuantum": "lambda VFCount : max( 32 // VFCount, 1)",
> +                        "VFPreemptionTimeout": "lambda VFCount : 128000 if (VFCount == 1) else max( 64000 // VFCount, 16000)"
> +                    }
> +                }
> +            },
> +            "Fixed_30fps_GPUTimeSlicing": {
> +                "GPUTimeSlicing": {
> +                    "ScheduleIfIdle": true,
> +                    "PFExecutionQuantum": 20,
> +                    "PFPreemptionTimeout": 20000,
> +                    "VFAttributes": {
> +                        "VFExecutionQuantum": "lambda VFCount : max( 32 // VFCount, 1)",
> +                        "VFPreemptionTimeout": "lambda VFCount : 128000 if (VFCount == 1) else max( 64000 // VFCount, 16000)"
> +                    }
> +                }
> +            },
> +            "Flexible_BurstableQoS_GPUTimeSlicing": {
> +                "GPUTimeSlicing": {
> +                    "ScheduleIfIdle": false,
> +                    "PFExecutionQuantum": 20,
> +                    "PFPreemptionTimeout": 20000,
> +                    "VFAttributes": {
> +                        "VFExecutionQuantum": "lambda VFCount : min((2000 // max(VFCount-1,1)*0.5, 50))",
> +                        "VFPreemptionTimeout": "lambda VFCount : (2000 // max(VFCount-1,1) - min((2000 // max(VFCount-1,1))*0.5, 50))*1000"
> +                    }
> +                }
> +            }
> +        }
> +    },
> +    "vGPUSecurity": {
> +        "Default": "Disabled",
> +        "Profile": {
> +            "Disabled": {
> +                "ResetAfterVfSwitch": false,
> +                "GuCSamplingPeriod": 0,
> +                "GuCThresholdCATError": 0,
> +                "GuCThresholdPageFault": 0,
> +                "GuCThresholdH2GStorm": 0,
> +                "GuCThresholdDbStorm": 0,
> +                "GuCThresholdGTIrqStorm": 0,
> +                "GuCThresholdEngineReset": 0
> +            }
> +        }
> +    }
> +}
> \ No newline at end of file
LGTM,
Reviewed-by: Marcin Bernatowicz <marcin.bernatowicz@linux.intel.com>

  reply	other threads:[~2024-12-05  9:51 UTC|newest]

Thread overview: 12+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2024-11-27 10:21 [PATCH i-g-t v3 0/4] vmtb: Introduce SR-IOV VM-level testing tool Adam Miszczak
2024-11-27 10:21 ` [PATCH i-g-t v3 1/4] tools/vmtb: VM Test Bench core Adam Miszczak
2024-11-27 10:22 ` [PATCH i-g-t v3 2/4] tools/vmtb: Basic SR-IOV tests Adam Miszczak
2024-11-27 10:22 ` [PATCH i-g-t v3 3/4] tools/vmtb: Test resources - vGPU profiles Adam Miszczak
2024-12-05  9:51   ` Bernatowicz, Marcin [this message]
2024-11-27 10:22 ` [PATCH i-g-t v3 4/4] tools/vmtb: Install VMTB with IGT build system Adam Miszczak
2024-11-27 10:58 ` ✓ Xe.CI.BAT: success for vmtb: Introduce SR-IOV VM-level testing tool (rev3) Patchwork
2024-11-27 11:18 ` ✓ i915.CI.BAT: " Patchwork
2024-11-27 13:38 ` ✗ i915.CI.Full: failure " Patchwork
2024-12-09 10:52   ` Adam Miszczak
2024-11-27 14:43 ` ✗ Xe.CI.Full: " Patchwork
2024-12-09 10:53   ` Adam Miszczak

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=37e9e22b-01e2-40ca-a761-b50980532be3@linux.intel.com \
    --to=marcin.bernatowicz@linux.intel.com \
    --cc=adam.miszczak@linux.intel.com \
    --cc=igt-dev@lists.freedesktop.org \
    --cc=kamil.konieczny@linux.intel.com \
    --cc=michal.wajdeczko@intel.com \
    --cc=pawel.sikora@intel.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.