This is a short post that explains how to write a high-performance matrix multiplication program on modern processors. In this tutorial I will use a single core of the Skylake-client CPU with AVX2, but the principles in this post also apply to other processors with different instruction sets (such as AVX512).
Matrix multiplication is a mathematical operation that defines the product of
| Latency Comparison Numbers (~2012) | |
| ---------------------------------- | |
| L1 cache reference 0.5 ns | |
| Branch mispredict 5 ns | |
| L2 cache reference 7 ns 14x L1 cache | |
| Mutex lock/unlock 25 ns | |
| Main memory reference 100 ns 20x L2 cache, 200x L1 cache | |
| Compress 1K bytes with Zippy 3,000 ns 3 us | |
| Send 1K bytes over 1 Gbps network 10,000 ns 10 us | |
| Read 4K randomly from SSD* 150,000 ns 150 us ~1GB/sec SSD |
I was at Amazon for about six and a half years, and now I've been at Google for that long. One thing that struck me immediately about the two companies -- an impression that has been reinforced almost daily -- is that Amazon does everything wrong, and Google does everything right. Sure, it's a sweeping generalization, but a surprisingly accurate one. It's pretty crazy. There are probably a hundred or even two hundred different ways you can compare the two companies, and Google is superior in all but three of them, if I recall correctly. I actually did a spreadsheet at one point but Legal wouldn't let me show it to anyone, even though recruiting loved it.
I mean, just to give you a very brief taste: Amazon's recruiting process is fundamentally flawed by having teams hire for themselves, so their hiring bar is incredibly inconsistent across teams, despite various efforts they've made to level it out. And their operations are a mess; they don't real
Each of these commands will run an ad hoc http static server in your current (or specified) directory, available at http://localhost:8000. Use this power wisely.
$ python -m SimpleHTTPServer 8000
| #!/bin/sh | |
| aws cloudformation deploy \ | |
| --template-file stack.yaml \ | |
| --stack-name edge-lambda-test \ | |
| --capabilities CAPABILITY_IAM \ | |
| --parameter-overrides Nonce=$RANDOM |
| Mute these words in your settings here: https://twitter.com/settings/muted_keywords | |
| ActivityTweet | |
| generic_activity_highlights | |
| generic_activity_momentsbreaking | |
| RankedOrganicTweet | |
| suggest_activity | |
| suggest_activity_feed | |
| suggest_activity_highlights | |
| suggest_activity_tweet |
| ################################################### | |
| ## | |
| ## Alertmanager YAML configuration for routing. | |
| ## | |
| ## Will route alerts with a code_owner label to the slack-code-owners receiver | |
| ## configured above, but will continue processing them to send to both a | |
| ## central Slack channel (slack-monitoring) and PagerDuty receivers | |
| ## (pd-warning and pd-critical) | |
| ## |
| ##TCP FLAGS## | |
| Unskilled Attackers Pester Real Security Folks | |
| ============================================== | |
| TCPDUMP FLAGS | |
| Unskilled = URG = (Not Displayed in Flag Field, Displayed elsewhere) | |
| Attackers = ACK = (Not Displayed in Flag Field, Displayed elsewhere) | |
| Pester = PSH = [P] (Push Data) | |
| Real = RST = [R] (Reset Connection) | |
| Security = SYN = [S] (Start Connection) |
