s

I have made entire HFT systems from nothing

[s @idoccor]

@ReedWhite18 No I never have. Kalshi actually failed to parse my passport the one time I tried to place a bet and I decided fuck that shit

[s @idoccor]

There are really 2 sides to HFT: the low latency aspect of placing orders quickly, and the bandwidth/HPC involved in parsing large amounts of historical data

Most systems I've replaced had a poor understanding of the computer science of one side or the other

[anon @anon597260576]

It’s hilarious because if your cousin making 70k a year in the IT department of BoA told you that story you’d be thinking wow their tech sucks ass

[s @idoccor]

@SCHIZO_FREQ Yeah but what if making fun of wanye is really funny?

[s @idoccor]

@kenwheeler Why do these things always have a eurotrash accent?

[s @idoccor]

Canada's going to send Angine de Poitrine to Eurovision and come in last place because that's the funniest thing that could happen.

[s @idoccor]

@yenwod_ But I didn’t read the whole article yet, I had to attend to my stupid wife and kids

[s @idoccor]

@yenwod_ Doing it this way lets you hand off normalized data to someone else asap (or use it yourself) while also building up a historical record

Also having normalized data lets you check integrity with some dataframe library

[s @idoccor]

@beniduboss mfw “I hope you and your family are well”

[s @idoccor]

@oleksoleksoleks @briskybop @Notbullnorbear @algoflows @CurbsideStocks @wesactual @Rsc77023274 @MisuaRaboki @Tim10337907 @MrnllMtt @KaV_2599 @UsamaButtar_ Yeah true

[s @idoccor]

@briskybop @Notbullnorbear @algoflows @CurbsideStocks @wesactual @Rsc77023274 @MisuaRaboki @Tim10337907 @MrnllMtt @KaV_2599 @UsamaButtar_ @oleksoleksoleks Gm

[s @idoccor]

@basedtensor More damning is that they hire web backend people from places like Google who proceed to fuck everything up because "I don't know how you guys do things in fiiiinance but at Google we ${moronic_idea}"

[s @idoccor]

@Dirnchen Usually both were busted. But with backtests in particular RAM consumption is typically an issue.

Or they use some framework like Spark that has nothing to do with trading but someone wanted it on their resume

[s @idoccor]

@liquiditygoblin https://t.co/R5oAauwvxG

[2Fast @2Fast_gg]

@anon597260576 @notsethbregman Wow this stripper really likes me

[s @idoccor]

@punished_daniel California already has 99 Ranch Market which is more or less the same thing

[s @idoccor]

@chatocv12 @blogagog I mmap compressed data to/from them and use them as if they were terrabytes of RAM. I’m not even sure if an EPYC board can hold 16TB of RDIMMs but it would be really expensive

[s @idoccor]

It’s stupid that consumer nvme drives all ship with a huge sticker on top of the memory chips but also run hot.

[s @idoccor]

@SalemIsTaken Yeah they're just paper mache Death From Above 1979. It's fine

[s @idoccor]

@blogagog Same, I use 4 of them in parallel for intensive market data processing and they run at 65-70C easily

[s @idoccor]

@DanielleFong @CattelainFrano1 Even cheap CPUs have laser etched branding on the heat spreader. Why should a $200 nvme be any different?

Especially when these things burn out like lightbulbs and/or throttle if you hammer them