Add a benchmarking app for Blazor Wasm #18015
Conversation
pranavkm
force-pushed
the
prkrishn/blazor-benchmarking
branch
from
efabacf to
bf5c512
Compare
|
|
||
| namespace Wasm.Performance.Driver | ||
| { | ||
| public partial class Program |
There was a problem hiding this comment.
Out of interest, does this code come from somewhere else, or is it newly-created for this PR?
I'm asking because it seems to contain a bunch of comments about working around issues and it's unusual to see that in the v1 of any particular bit of code.
There was a problem hiding this comment.
It makes me wonder if we're duplicating knowledge across projects through copy-paste and if there are opportunities to centralise this infrastructure somehow.
|
|
||
| var psi = new ProcessStartInfo | ||
| { | ||
| FileName = "/opt/bin/start-selenium-standalone.sh", |
There was a problem hiding this comment.
???
Is this only intended to run within a particular Docker image? If so perhaps this could be clarified via the naming of the project or something.
There was a problem hiding this comment.
I don't really understand why we need a Linux-specific version of "start selenium" - don't we already have a cross-plat version of this logic in the E2E test infrastructure?
| To run the benchmark app in the Benchmark server, run | ||
|
|
||
| ``` | ||
| dotnet run -- --config \aspnetcore\src\Components\benchmarkapps\Wasm.Performance\benchmarks.compose.json --services.blazorwasmbenchmark.endpoints <BenchmarkServerUri> |
There was a problem hiding this comment.
Should this command be run inside the Docker container, or on the host machine?
I'm guessing within the container, but then I don't follow why the path has backslashes.
|
This is very interesting stuff and is probably all great. There's quite a lot going on in here that feels very different to the infrastructure we have had elsewhere in this repo (mainly, the use of Docker) so I'd ideally like to have a meeting to discuss. Would you be willing to do a demo of this and explain the reasoning behind the design? For example I'd like to see the steps involved in running the benchmarks locally during development. I'm not even clear right now why this should involve Docker at all during automated runs, though I'm totally willing to believe it's for a valid reason. |
pranavkm
force-pushed
the
prkrishn/blazor-benchmarking
branch
8 times, most recently
from
5625263 to
d1d9548
Compare
pranavkm
force-pushed
the
prkrishn/blazor-benchmarking
branch
2 times, most recently
from
3816064 to
e57c046
Compare
pranavkm
force-pushed
the
prkrishn/blazor-benchmarking
branch
from
e57c046 to
1e62df1
Compare
|
@pranavkm Glad to see this is merged! At some point, will you be able to let us know if there's some URL we can visit to see the benchmarking results? It would be amazing to see the numbers for each commit so that we can assess the perf impact of major changes (such as upgrading to new versions of Mono WebAssembly). |
No description provided.