VEVerilog · Beginners
Community homeHot discussions Latest posts Best of Knowledge base Resources About

Verilog · Beginners · The edge cases the verilog-newbie docs never spell out

304
VEr/verilog-newbie·posted by mike_xu·3 days agoTutorialPinned

The edge cases the verilog-newbie docs never spell out

Most verilog-newbie articles stop at "how to use it" and never cover "when not to use it". This is an attempt at the second half.

The first thing was to collapse the variables. We were changing config and upgrading the version at the same time, and afterwards nobody could say which change caused what. We rolled back to moving one variable at a time, re-ran three times, and only then did the curve settle. Tedious, but not skippable.

Self-host — full control44%
Managed service — less work28%
Hybrid: core self-hosted17%
Not decided yet11%

517 votes total

61 comments

61 comments

M
Zzhu_zongOP·1 hour ago

I just read the verilog-newbie source — the author actually explains the reasoning in a comment, roughly "so that it degrades into predictable behaviour in extreme cases".

291
Cchen_dev·2 days ago

This matches what we see in production. We only hit it past 3k QPS; the earlier load tests showed nothing — the test traffic was too clean, with no long-tail requests.

343
Sswoole_leeOP·2 days ago

A question: what changes in a container with a 512Mi memory limit? That is how we run it in production.

278
Rran_bo·2 days ago

I see point 3 differently. The trade-off depends on your read/write ratio: read-heavy with little writing means caching actually widens the inconsistency window.

125
Ttang_hao·28 minutes ago

Has anyone run a controlled experiment? I did, reducing it to a single variable, and the difference was 4% — within noise. So I suspect the main cause is something else.

119
Llinlin·just now

A question: what changes in a container with a 512Mi memory limit? That is how we run it in production.

109
Bbob_chen·2 days agoedited

Worth learning from this debugging approach. We went straight at the logs and took a much longer route.

107
Sswoole_lee·2 days ago

One counter-example: below verilog-newbie 7.4 the semantics of that code are different, so do not copy it verbatim. We got burned in staging and rolled back once.

94
Mmike_xuOP·3 minutes ago

This matches what we see in production. We only hit it past 3k QPS; the earlier load tests showed nothing — the test traffic was too clean, with no long-tail requests.

24
Cchen_dev·yesterday

This is not a verilog-newbie problem, it is a usage problem. The docs say this API is not thread-safe and you must lock around it yourself.

78
Lli_ming·1 hour ago

There is actually a simpler fix that needs no architecture change: move this check up to the gateway and the problem disappears. The cost is one extra lookup at the gateway.

202
Sswoole_lee·2 days ago

This matches what we see in production. We only hit it past 3k QPS; the earlier load tests showed nothing — the test traffic was too clean, with no long-tail requests.

135
Bbob_chenOP·just now

Saved. I am reworking this area this week — this saves a lot of wrong turns.

17
Hhuang_ke·5 hours ago

This is not a verilog-newbie problem, it is a usage problem. The docs say this API is not thread-safe and you must lock around it yourself.

412
Aalice_devOP·just now

Thanks for sharing real numbers — far more useful than the articles that only cover concepts.

122
Nnikic·2 days ago

One counter-example: below verilog-newbie 7.4 the semantics of that code are different, so do not copy it verbatim. We got burned in staging and rolled back once.

399
Wwinter·2 days ago

A question: what changes in a container with a 512Mi memory limit? That is how we run it in production.

397
Bbob_chenOP·2 days ago

Has anyone run a controlled experiment? I did, reducing it to a single variable, and the difference was 4% — within noise. So I suspect the main cause is something else.

78
Sswoole_lee·1 hour ago

Can you give a minimal reproduction? I ran it locally for ten minutes and could not reproduce on macOS with the latest version.

306
Wwinter·2 days ago

Agreeing with the above. One addition: with this option enabled the GC count in your metrics doubles, so adjust the alert threshold at the same time or it will keep firing.

209
Kkernel_panic·2 hours agoedited

We have run this in production for two years without hitting it. That said, we never reached this scale, so our experience is not really evidence here.

4
Ttang_haoMod·2 days ago

Agreeing with the above. One addition: with this option enabled the GC count in your metrics doubles, so adjust the alert threshold at the same time or it will keep firing.

486
Lli_ming·2 days ago

This is not a verilog-newbie problem, it is a usage problem. The docs say this API is not thread-safe and you must lock around it yourself.

255
Llinlin·1 hour agoedited

Sharing our numbers, 8 cores 16GB, same scenario:

| Concurrency | P50 | P99 |
|---|---|---|
| 200 | 12ms | 88ms |
| 500 | 31ms | 340ms |

P99 clearly collapses at 500 concurrency, which lines up with your knee point.

286
Rran_bo·1 hour ago

I just read the verilog-newbie source — the author actually explains the reasoning in a comment, roughly "so that it degrades into predictable behaviour in extreme cases".

356
Zzhou_yi·5 hours ago

There is actually a simpler fix that needs no architecture change: move this check up to the gateway and the problem disappears. The cost is one extra lookup at the gateway.

444
Cchen_dev·2 days agoeditedLevel 6

One counter-example: below verilog-newbie 7.4 the semantics of that code are different, so do not copy it verbatim. We got burned in staging and rolled back once.

400
Wwinter·2 days agoLevel 6

This is not a verilog-newbie problem, it is a usage problem. The docs say this API is not thread-safe and you must lock around it yourself.

272
Bbob_chen·2 days agoLevel 6

Has anyone run a controlled experiment? I did, reducing it to a single variable, and the difference was 4% — within noise. So I suspect the main cause is something else.

133
Sswoole_lee·12 minutes ago

Worth learning from this debugging approach. We went straight at the logs and took a much longer route.

393
Cchen_dev·2 days agoLevel 6

Saved. I am reworking this area this week — this saves a lot of wrong turns.

3
Mmike_xu·2 days ago

Thanks for sharing real numbers — far more useful than the articles that only cover concepts.

66
Lli_ming·2 days agoLevel 6

We have run this in production for two years without hitting it. That said, we never reached this scale, so our experience is not really evidence here.

26
Hhuang_ke·2 days ago

Can you give a minimal reproduction? I ran it locally for ten minutes and could not reproduce on macOS with the latest version.

3
Bbob_chen·2 days agoedited

Sharing our numbers, 8 cores 16GB, same scenario:

| Concurrency | P50 | P99 |
|---|---|---|
| 200 | 12ms | 88ms |
| 500 | 31ms | 340ms |

P99 clearly collapses at 500 concurrency, which lines up with your knee point.

146
Ttang_hao·12 minutes ago

We have run this in production for two years without hitting it. That said, we never reached this scale, so our experience is not really evidence here.

4
Aalice_dev·5 hours ago

I see point 3 differently. The trade-off depends on your read/write ratio: read-heavy with little writing means caching actually widens the inconsistency window.

202
Sslow_query·2 days ago

Sharing our numbers, 8 cores 16GB, same scenario:

| Concurrency | P50 | P99 |
|---|---|---|
| 200 | 12ms | 88ms |
| 500 | 31ms | 340ms |

P99 clearly collapses at 500 concurrency, which lines up with your knee point.

2
Oops_wang·5 hours ago

Saved. I am reworking this area this week — this saves a lot of wrong turns.

431
WwinterOP·2 days ago

Saved. I am reworking this area this week — this saves a lot of wrong turns.

219
Kkite·2 days agoLevel 6

I just read the verilog-newbie source — the author actually explains the reasoning in a comment, roughly "so that it degrades into predictable behaviour in extreme cases".

347
Ddev_zhou·2 days agoLevel 6

Can you give a minimal reproduction? I ran it locally for ten minutes and could not reproduce on macOS with the latest version.

19
Mmike_xu·2 days agoedited

Thanks for sharing real numbers — far more useful than the articles that only cover concepts.

121
Sswoole_lee·3 minutes ago

Can you give a minimal reproduction? I ran it locally for ten minutes and could not reproduce on macOS with the latest version.

75
Oops_wang·2 days ago

I see point 3 differently. The trade-off depends on your read/write ratio: read-heavy with little writing means caching actually widens the inconsistency window.

38
Zzhou_yiOP·2 days ago

Can you give a minimal reproduction? I ran it locally for ten minutes and could not reproduce on macOS with the latest version.

6
Aalice_devOP·2 days agoedited

I just read the verilog-newbie source — the author actually explains the reasoning in a comment, roughly "so that it degrades into predictable behaviour in extreme cases".

2
Ddev_zhou·28 minutes ago

I see point 3 differently. The trade-off depends on your read/write ratio: read-heavy with little writing means caching actually widens the inconsistency window.

34
Aalice_dev·2 days ago

This is not a verilog-newbie problem, it is a usage problem. The docs say this API is not thread-safe and you must lock around it yourself.

413
Oops_wang·2 days ago

Agreeing with the above. One addition: with this option enabled the GC count in your metrics doubles, so adjust the alert threshold at the same time or it will keep firing.

151
Rrase·2 days ago

Has anyone run a controlled experiment? I did, reducing it to a single variable, and the difference was 4% — within noise. So I suspect the main cause is something else.

8
Lli_ming·2 days ago

We have run this in production for two years without hitting it. That said, we never reached this scale, so our experience is not really evidence here.

139
Sslow_query·yesterday

Worth learning from this debugging approach. We went straight at the logs and took a much longer route.

16
Kkernel_panic·2 days agoedited

Worth learning from this debugging approach. We went straight at the logs and took a much longer route.

16
Zzhou_yi·yesterday

One counter-example: below verilog-newbie 7.4 the semantics of that code are different, so do not copy it verbatim. We got burned in staging and rolled back once.

8
Sslow_query·yesterday

Thanks for sharing real numbers — far more useful than the articles that only cover concepts.

5
Zzhou_yi·12 minutes ago

Agreeing with the above. One addition: with this option enabled the GC count in your metrics doubles, so adjust the alert threshold at the same time or it will keep firing.

7
Zzhu_zong·3 minutes ago

There is actually a simpler fix that needs no architecture change: move this check up to the gateway and the problem disappears. The cost is one extra lookup at the gateway.

29
Kkernel_panic·2 days ago

Sharing our numbers, 8 cores 16GB, same scenario:

| Concurrency | P50 | P99 |
|---|---|---|
| 200 | 12ms | 88ms |
| 500 | 31ms | 340ms |

P99 clearly collapses at 500 concurrency, which lines up with your knee point.

15
Sswoole_lee·2 days ago

A question: what changes in a container with a 512Mi memory limit? That is how we run it in production.

6
Rran_bo·28 minutes ago

This matches what we see in production. We only hit it past 3k QPS; the earlier load tests showed nothing — the test traffic was too clean, with no long-tail requests.

1

This is the post detail page /en/c/verilog-newbie/post/p0. Posts and comments are generated deterministically from a seeded PRNG, so the same post always renders the same content and the link can be shared, reloaded and indexed. In production this page reads MySQL for the post, Redis for hot-post caching, and fetches the whole comment tree in a single query on the path column.

See the database schema →