Merge pull request #360 from sunya-ch/machine-sponsor

feat: add call for machines

docs/sponsor/call-for-machines.md

@@ -0,0 +1,78 @@
+# Call for Machines
+
+## Expected Usage
+
+- Periodic report: run training and validation workflow once per week and each run can spend upto 1 hour depends on the machine processing speed.
+  The frequency and period of running could be changed with notice and will be applied only with approval.
+
+- Release execution: run training, validation, and export workflow on each Kepler release.
+
+## Type of Machines
+
+As mentioned in several research works, the power consumption behavior can be varied by several factors.
+Using the right power model to predict power consumption on the machine that has no power meter is critical the precision of the reported values.
+
+We are seeking machines with the following characteristics, commonly used in cloud computing environments but not limited to:
+
+### Processor
+- Intel: Cascade Lake, Skylake (or newer)
+- Ampere: Altra, AmpereOne (or newer)
+- AMD: Opteron, EPYC (or newer)
+
+### Accelerator Chips
+
+- NVML-supported GPU
+
+    |Architecture|Microarchitecture|Supported GPUs|
+    |---|---|---|
+    |Kepler|GK104, GK106, GK107|GeForce GTX 600 series (e.g., GTX 660, GTX 670), GTX 700 series|
+    |Maxwell|GM204, GM206|GeForce GTX 900 series (e.g., GTX 970, GTX 980)|
+    |Pascal|GP100, GP102|GeForce GTX 10 series (e.g., GTX 1080, GTX 1070), Titan X|
+    |Volta|GV100|Tesla V100, Titan V|
+    |Turing|TU102, TU104|GeForce RTX 20 series (e.g., RTX 2080 Ti, RTX 2070)|
+    |Ampere|GA100|A100, A40, RTX 30 series (e.g., RTX 3090)|
+    |Ada Lovelace|AD102|RTX 40 series (e.g., RTX 4090)|
+    |Hopper|GH100|H100|
+
+
+- DCGM-supported GPU
+
+    |Architecture|Supported GPUs|
+    |---|---|
+    |Kepler|Tesla K80 and newer|
+    |Maxwell|Tesla M10, Tesla M60|
+    |Pascal|Tesla P100, Tesla P40, Tesla P4|
+    |Volta|Tesla V100|
+    |Turing|T4, Quadro RTX 4000, RTX A4000|
+    |Ampere|A100, A40, A30, A10, A2, RTX A6000, RTX A5000|
+    |Hopper|H100, H200|
+
+- HLML-supported Intel Gaudi
+
+> Please note that power source modules for other accelerator cards, such as AMD GPUs, Google TPUs, and IBM AIUs, are not yet supported by Kepler.
+> Contributions to support these devices are always welcome within the Kepler community.
+
+### Power meters
+- rapl
+- powerclamp
+- amd-pstate
+- acpi
+- hmc
+- redfish
+- nvml (nvidia)
+- hlml (habana)
+- dcgm
+
+### Available machines on CI
+
+List of currently available bare metal machines for power model training and validation CI pipeline in [kepler-metal-ci](https://github.com/sustainable-computing-io/kepler-metal-ci/tree/main).
+
+|processor|cores|chips|memory|power meters|source|sponsor|key contact
+|---|---|---|----|---|---|---|---|
+|intel_xeon_e_2278g| 16 | 1 | - | RAPL (package, core, dram)|equinix c3.small.x86|CNCF|@rootfs|
+|intel_xeon_platinum_8259cl| 96 | 2 | 377| RAPL (package, dram)|ec2 m5.metal|RedHat|@rootfs|
+|intel_xeon_e5_2686v4| 72 | 2 | 503| RAPL (package, dram)|ec2 i3.metal|RedHat|@rootfs|
+|intel_xeon_platinum_8275cl| 96 | 2 | 188| RAPL (package, dram)|ec2 c5.metal|RedHat|@rootfs|
+|intel_xeon_platinum_8259cl| 96 | 2 | 755| RAPL (package, dram)|ec2 r5.metal|RedHat|@rootfs|
+|intel_xeon_platinum_8252c| 48 | 2 | 188 | RAPL (package, dram)|ec2 m5zn.metal|RedHat|@rootfs|
+|intel_xeon_platinum_8488c| 96 | 1 | 377 | RAPL (package, dram)|ec2 m7i.metal-24xl|RedHat|@rootfs|