Adding ecmascript simd tests.

Adding ecmascript simd tests.

Relying on the polyfill for now.

BUG=v8:4124
LOG=N
R=bbudge@chromium.org,machenbach@chromium.org
TEST=trybots

Committed: https://crrev.com/8dfec46e38853ab7503662eb17f8e5475663ec2b
Cr-Commit-Position: refs/heads/master@{#28470}

[Michael Achenbach, 2015-05-19 07:17:48]

https://codereview.chromium.org/1147743004/diff/1/test/simdjs/harness-adapt.js
File test/simdjs/harness-adapt.js (right):

https://codereview.chromium.org/1147743004/diff/1/test/simdjs/harness-adapt.j...
test/simdjs/harness-adapt.js:47: _oldLoad('test/simdjs/data/src/benchmarks/' +
filename);
Did you test that? Not sure if it'll work as the working dir for benchmark
execution is the same as the json file of the benchmark or relatively deeper (as
specified in the file).

We can also fix that later when landing the benchmark stuff.

https://codereview.chromium.org/1147743004/diff/1/test/simdjs/testcfg.py
File test/simdjs/testcfg.py (right):

https://codereview.chromium.org/1147743004/diff/1/test/simdjs/testcfg.py#newc...
test/simdjs/testcfg.py:1: # Copyright 2015 the V8 project authors. All rights
reserved.
Please use shorter licence headers as in
https://code.google.com/p/chromium/codesearch#chromium/src/v8/test/unittests/...

Also in the other files.

https://codereview.chromium.org/1147743004/diff/1/test/simdjs/testcfg.py#newc...
test/simdjs/testcfg.py:56: tests = [
This wraps another runner? Is there a way to get the list of test cases and then
make the runner execute separately for each test case (passing in the test
name)? To wrap a runner has several disadvantages:
1. It doesn't use our runner's multiprocessing
2. Any set of test failures will appear as one failure of that runner.
3. The status file can't be use to manipulate specific test runs.

How many atomic tests are called by the runner and how long does that take?

[bbudge, 2015-05-19 08:36:02]

https://codereview.chromium.org/1147743004/diff/1/test/simdjs/harness-adapt.js
File test/simdjs/harness-adapt.js (right):

https://codereview.chromium.org/1147743004/diff/1/test/simdjs/harness-adapt.j...
test/simdjs/harness-adapt.js:29: 
"use strict"; ?

[Michael Achenbach, 2015-05-19 09:34:48]

https://codereview.chromium.org/1147743004/diff/1/test/simdjs/harness-adapt.js
File test/simdjs/harness-adapt.js (right):

https://codereview.chromium.org/1147743004/diff/1/test/simdjs/harness-adapt.j...
test/simdjs/harness-adapt.js:47: _oldLoad('test/simdjs/data/src/benchmarks/' +
filename);
On 2015/05/19 07:17:48, Michael Achenbach wrote:
> Did you test that? Not sure if it'll work as the working dir for benchmark
> execution is the same as the json file of the benchmark or relatively deeper
(as
> specified in the file).
> 
> We can also fix that later when landing the benchmark stuff.

Just ignore this. I was confused. I assume when running the benchmarks with the
run_perf we don't use this file at all...

[bradn, 2015-05-19 09:37:53]

bradnelson@google.com changed reviewers:
+ bradnelson@google.com

[bradn, 2015-05-19 09:37:53]

https://codereview.chromium.org/1147743004/diff/1/test/simdjs/harness-adapt.js
File test/simdjs/harness-adapt.js (right):

https://codereview.chromium.org/1147743004/diff/1/test/simdjs/harness-adapt.j...
test/simdjs/harness-adapt.js:29: 
On 2015/05/19 08:36:01, bbudge (OOO week of 5-11) wrote:
> "use strict"; ?

Done.

https://codereview.chromium.org/1147743004/diff/1/test/simdjs/harness-adapt.j...
test/simdjs/harness-adapt.js:47: _oldLoad('test/simdjs/data/src/benchmarks/' +
filename);
On 2015/05/19 07:17:48, Michael Achenbach wrote:
> Did you test that? Not sure if it'll work as the working dir for benchmark
> execution is the same as the json file of the benchmark or relatively deeper
(as
> specified in the file).
> 
> We can also fix that later when landing the benchmark stuff.

So the benchmarks or these tests run ok if you change the cwd to where the test
is.
This also works to run both, though the benchmarks are slow.

https://codereview.chromium.org/1147743004/diff/1/test/simdjs/testcfg.py
File test/simdjs/testcfg.py (right):

https://codereview.chromium.org/1147743004/diff/1/test/simdjs/testcfg.py#newc...
test/simdjs/testcfg.py:1: # Copyright 2015 the V8 project authors. All rights
reserved.
On 2015/05/19 07:17:48, Michael Achenbach wrote:
> Please use shorter licence headers as in
>
https://code.google.com/p/chromium/codesearch#chromium/src/v8/test/unittests/...
> 
> Also in the other files.

Done.

https://codereview.chromium.org/1147743004/diff/1/test/simdjs/testcfg.py#newc...
test/simdjs/testcfg.py:56: tests = [
On 2015/05/19 07:17:48, Michael Achenbach wrote:
> This wraps another runner? Is there a way to get the list of test cases and
then
> make the runner execute separately for each test case (passing in the test
> name)? To wrap a runner has several disadvantages:
> 1. It doesn't use our runner's multiprocessing
> 2. Any set of test failures will appear as one failure of that runner.
> 3. The status file can't be use to manipulate specific test runs.
> 
> How many atomic tests are called by the runner and how long does that take?

Ah, good point.
I've changed it round to expose the individual ones.

[Michael Achenbach, 2015-05-19 09:40:56]

lgtm

[bradn, 2015-05-19 09:48:42]

The CQ bit was checked by bradnelson@google.com

[commit-bot: I haz the power, 2015-05-19 09:48:48]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1147743004/40001

[commit-bot: I haz the power, 2015-05-19 10:14:54]

Committed patchset #3 (id:40001)

[commit-bot: I haz the power, 2015-05-19 10:15:14]

Patchset 3 (id:??) landed as
https://crrev.com/8dfec46e38853ab7503662eb17f8e5475663ec2b
Cr-Commit-Position: refs/heads/master@{#28470}