Robert Bindar

@choldgraf I don't know if there is a documented limit on what GraphQL wants/is able to return, I'll research more. One year is what I discovered works for the org/repo...

[Update] Resource throttling seems to be documented here: https://developer.github.com/v4/guides/resource-limitations/

> It would be better if this function were extensible to allow integral types. This is an interface issue, not a request for such code in the current PR. One...

@eric-hughes-tiledb updated the code based on what we last discussed. I have some comments/questions here: 1. I adjusted generated vectors from 2^24 as we discussed to 2^22 as the test...

@choldgraf many thanks for the quick answer! :-) On Jupyter activity analysis, Very cool work! Hope you won't mind if I might try to find inspiration in the metrics and...

Reproduced on i7i.24xlarge, I'm getting 8186MB/s read and write bandwidths with a default iotune master build. 1.1M IOPS write, very similar to what @syuu1228 got.

> Low write iops numbers can be because iotune mis-detects block size. From sysfs it can read, say, 1k, but device "prefers" 4k and performs read-modify-write on it. There's an...

> > > Low write iops numbers can be because iotune mis-detects block size. From sysfs it can read, say, 1k, but device "prefers" 4k and performs read-modify-write on it....

@mykaul I've been running some experiments with Pavel on i7i.12xlarge. I actually started running on a 24xlarge machine, but after the first evening, I couldn't get one available anymore. Here's...

> Do we suspect the bottleneck is somewhere with: > > 1. iotune itself?! I doubt it for bandwidth measurements. the problems with iotune are: - it fails to detect...