0
Planned

Reply doesn't work correctly after Edit/Delete

zybexXL 1 year ago in Website updated by Capt Jack Daniels 1 year ago 6

Hi, I found an issue on the article comment system:

1. Publish a comment

2. Click Edit/Delete, but don't change anything (you changed your mind)

3. Click Reply of a different comment, type some text

4. There's an "Update" button instead of "Publish" (bug 1)

5. Click "Update" - your original comment is replaced by this reply (bug 2, related)

This causes the original comment to be lost and the reply to be posted instead.

Under review

Thank you for notifying us about this inconvenient issue, we will try to fix as soon as possible. 

In a meantime, you can click on close (X) button at the top of comment when it is in Edit mode to avoid loosing the comments.

Thank you.

Is it possible for you to find my previous comment which was lost and replaced with this one?

https://seekingalpha.com/article/4284981-amd-rome-launch-marks-watershed-moment-companys-history#comment-82569559

If you can revert the edit please go ahead.

If this is the right one, please feel free to copy paste it back in

I fully agree with the title. Proper Epyc Rome reviews: https://www.servethehome.com/amd-epyc-7002-series-rome-delivers-a-knockout/ https://www.phoronix.com/scan.php?page=article&item=amd-epyc-7502-7742&num=1 You just need to look at the Performance per Dollar and Performance per Watt charts to understand why this CPU is a massive improvement over Intel's solutions. For datacenters, TCO is critical. @EnerTuition - some corrections regarding performance: "Some workloads, especially AI related ones, Intel leads due support for Google bfloat16 instructions. Some of Intel’s chips are also advantaged due to high performance AVX-512 implementation which AMD does not have." Bfloat16 is not yet available on Intel/AMD CPUs. Intel will have on the Cooper Lake Xeons, due in 2020 (don't hold your breath). By then Zen3/Milan will also be out, and it may also bring BFloat16 and AVX512. Ice Lake (Intel 10nm) might not even have BFloat16, which would be strange since it comes after Cooper Lake: https://www.anandtech.com/show/14179/intel-manual-updates-bfloat16-for-cooper-lake-xeon-scalable-only Epyc Rome doesn't have AVX512, but benchmarks show that Rome's AVX2 (256bit) is as fast as Intel's AVX512 - see GROMACS benchmark: https://www.servethehome.com/amd-epyc-7002-series-rome-delivers-a-knockout/8/ The only benchmarks where Intel still has an advantage are the single threaded ones due to higher boost clock speeds. However, no one buys a 64-core CPU to then just use one of the cores. When using many cores, clock speed drops and the advantage goes [massively] to AMD. Another benchmark where Intel wins is MKL-DNN... which is developed by Intel. Obviously only optimized for Intel, and using generic code for non-Intel CPUs. Intel also wins on MySQL/MariaDB - AMD says it's investigating, this one is abnormal.

Thank you Jacob, that was it :)

When I tried to correct a  misspelled word  the nice moderators simply deleted my post instead.