Bugzilla

Per comment1, the GPU which the reporter uses is `Intel(R) HD Graphics 620`. On my Dell Xp15 machine, my GPU is `Intel(R) UHD Graphics 630` which is very similar with 620, based on this [comparison](https://www.notebookcheck.net/UHD-Graphics-620-vs-UHD-Graphics-630_8098_8126.247598.0.html). However, I didn't encounter any dropped frame on the same [video](https://www.notebookcheck.net/UHD-Graphics-620-vs-UHD-Graphics-630_8098_8126.247598.0.html) where the reporter saw around 6~8% dropped frames. Here is my [profiled result](https://share.firefox.dev/2YETlDq), the decoding took very short time and `NtGdiDdDDIWaitForSynchronizationObjectFromCpu` only took less than 1% time on the GPU process.

Based on this [comment](https://github.com/microsoft/angle/issues/111#issuecomment-279536657), MS people said the problem was caused by ANGLE, and the reporter uses `ANGLE 2.1.14225` and mine is `ANGLE 2.1.15727`. And the [lastest comment](https://github.com/microsoft/angle/issues/111#issuecomment-533701795) (2019 Sep) showed that the problem was still in the ANGLE master at that time. As I don't know much about ANGLE, is  that something we can control? Or we need to ask users to update their ANGLE version?

Thank you.

Per comment1, the GPU which the reporter uses is `Intel(R) HD Graphics 620`. On my Dell Xp15 machine, my GPU is `Intel(R) UHD Graphics 630` which is very similar with 620, based on this [comparison](https://www.notebookcheck.net/UHD-Graphics-620-vs-UHD-Graphics-630_8098_8126.247598.0.html). However, I didn't encounter any dropped frame on the same [video](https://www.notebookcheck.net/UHD-Graphics-620-vs-UHD-Graphics-630_8098_8126.247598.0.html) where the reporter saw around 6~8% dropped frames. Here is my [profiled result](https://share.firefox.dev/2YETlDq), the decoding took very short time and `NtGdiDdDDIWaitForSynchronizationObjectFromCpu` only took less than 1% time on the GPU process.

Based on this [comment](https://github.com/microsoft/angle/issues/111#issuecomment-279536657), MS people said the problem was caused by ANGLE, and the reporter uses `ANGLE 2.1.14225` and mine is `ANGLE 2.1.15727`. And the [lastest comment](https://github.com/microsoft/angle/issues/111#issuecomment-533701795) (2019 Sep) showed that the problem was still in the ANGLE master at that time, but it probably got fixed in the some version already. As I don't know much about ANGLE, is  that something we can control? Or we need to ask users to update their ANGLE version?

Thank you.