Google Groups no longer supports new Usenet posts or subscriptions. Historical content remains viewable.
Dismiss

Re: High Frequency Trading/C++

186 views
Skip to first unread message
Message has been deleted

Paavo Helde

unread,
Sep 14, 2014, 3:52:48 PM9/14/14
to
Robert Hutchings <rm.hut...@gmail.com> wrote in
news:cb98adef-cf94-426e...@googlegroups.com:

> What C++ skills are most relevant to HFT? I think multithreading and
> ultra-low-latency would be at the top. Any others?

These are not exactly "C++ skills". And from the first google hits I read:

"many UHFT firms have moved towards custom hardware"

so I guess THDL++ (http://visualhdl.sysprogs.org/thdlpp/) would be the
closest connection with C++.

Honestly, if you need to ask then you probably aren't qualified for the job
anyway ;)

Cheers
Paavo


Message has been deleted

David Brown

unread,
Sep 15, 2014, 10:57:11 AM9/15/14
to
On 15/09/14 16:01, Robert Hutchings wrote:
> Was it really necessary to be so condescending?
>

High frequency trading is seriously specialised development, with a mix
of hardware, software, networking, economics, and other fields. Asking
in a group like this for the "most relevant C++ skills" is a bit like
asking for what painting skills are most relevant to designing jumbo jets.

So Paavo was simply trying to make it clear to you that you are not
ready to be writing HFT trading code in C++.

(Of course, that doesn't mean you could not get a job working for an HFT
company, and learning as you go along.)

Jorgen Grahn

unread,
Sep 15, 2014, 12:00:54 PM9/15/14
to
On Mon, 2014-09-15, David Brown wrote:
> On 15/09/14 16:01, Robert Hutchings wrote:
>> Was it really necessary to be so condescending?
>>
>
> High frequency trading is seriously specialised development, with a mix
> of hardware, software, networking, economics, and other fields. Asking
> in a group like this for the "most relevant C++ skills" is a bit like
> asking for what painting skills are most relevant to designing jumbo jets.
>
> So Paavo was simply trying to make it clear to you that you are not
> ready to be writing HFT trading code in C++.

I read the original question as more of "I'm curious what role C++ has
in HFT" than "tell me how to make lots of money doing HTF in C++".
From that angle, Paavo's answer could have been more ... sensitive,
perhaps. Despite the smiley.

/Jorgen

--
// Jorgen Grahn <grahn@ Oo o. . .
\X/ snipabacken.se> O o .
Message has been deleted

Scott Lurndal

unread,
Sep 15, 2014, 12:52:39 PM9/15/14
to
There really isn't any tie between HFT and C++. C++ is a tool, and
as a tool, it can be applied to many problems.

Modern HFT relies more on hardware than software (e.g. low-latency
interconnects to the exchanges, such that most brokerages and investment
banks colocate equipment in the exchange, and the fiber links between
the exchange systems and colocated hosts are intentionally designed
to be equal in length (so that one colocated host doesn't get a
feed a few nanoseconds ahead of another)).

Consider, for example, criteria for interconnect links:

Transport Technology Switch/Endpoint Encoding/Decoding/Equalization latency
-------------------- ---------------------------
10G (SFP to Fibre) 300ns
10G-BaseT (Copper) 2500 ns

That 2200ns is circa 6000 instructions on a 3Ghz processor, depending
on IPC. Can give a trader quite an advantage, if for a competitor
every switch hop/host ingress or egress adds 2200ns latency.

The ratio is roughly the same for 40G and 100G links.

Note that this eliminates TCP as a viable transport - most use UDP and
take care to ensure that there aren't dropped packets.

Algorithmic processing is often offloaded to custom coprocessors
(generally Field-Programmable Gate Arrays, which are programmed in
Verilog/VHDL).

One should be comfortable with B-S[*] and derivitives to operate in such environments.

[*] http://en.wikipedia.org/wiki/Black%E2%80%93Scholes_model

Given the low-latency requirements, C++ may not be the best choice anyway.

Victor Bazarov

unread,
Sep 15, 2014, 2:01:40 PM9/15/14
to
On 9/15/2014 12:00 PM, Jorgen Grahn wrote:
> [..] Paavo's answer could have been more ... sensitive,
> perhaps. Despite the smiley.

Honestly, how is one going to make it in the cut-throat world of modern
automated trading if he can't even take a newsgroup jab, especially one
accompanied by a smiley?..

"Down with newsgroup bullying!!!" -- I can just see CNN spending hours a
day chewing fat on that... <sigh>

V
--
I do not respond to top-posted replies, please don't ask

Victor Bazarov

unread,
Sep 15, 2014, 2:02:53 PM9/15/14
to
On 9/15/2014 12:05 PM, Robert Hutchings wrote:
> Wow. I daresay I won't be posting in this group of super-experts again. [..]

Your choice, Robert.
Message has been deleted
Message has been deleted

Victor Bazarov

unread,
Sep 15, 2014, 2:48:06 PM9/15/14
to
On 9/15/2014 2:41 PM, Robert Hutchings wrote:
> Have you taken all your meds yet Victor? Or are you always this hostile?

So you are back? Could stay away long, could you?

Glad to have you, Robert! I see you're learning the art of newsgroup
flame war. When no technical argumentation left at your disposal,
attack on the personal level! Keep it up!
Message has been deleted
Message has been deleted

Victor Bazarov

unread,
Sep 15, 2014, 2:55:36 PM9/15/14
to
On 9/15/2014 2:52 PM, Robert Hutchings wrote:
> Okay, I WILL keep it up!

Glad you're all we expected you to be, Robert!
Message has been deleted

Victor Bazarov

unread,
Sep 15, 2014, 3:00:26 PM9/15/14
to
On 9/15/2014 2:54 PM, Robert Hutchings wrote:
> Oh Victor. <SIGN> ..what a shame that you are on this forum.

I guess you meant "<SIGH>"...

Shame? Care to elaborate? Do *you* feel shame? Or should somebody
else feel shame because they are somehow responsible? And for what,
exactly?

Ian Collins

unread,
Sep 15, 2014, 3:02:28 PM9/15/14
to
Robert Hutchings wrote:
> Uh, are you English-challenged sir? I thought so. Oh well, I guess I'll go ahead and "top-post" something...
>
Who or what are you replying to? At least take the trouble to learn how
to post if you are gong to attempt a smart arse reply.

--
Ian Collins

Victor Bazarov

unread,
Sep 15, 2014, 3:02:45 PM9/15/14
to
On 9/15/2014 2:56 PM, Robert Hutchings wrote:
> Uh, are you English-challenged sir? I thought so. Oh well, I guess I'll go ahead and "top-post" something...

Please. As soon as you figure out how to actually do that...

Hint: without quoting it's impossible to top-post. If you need more
explanation, just ask.
Message has been deleted
Message has been deleted
Message has been deleted

Ian Collins

unread,
Sep 15, 2014, 3:18:13 PM9/15/14
to
Robert Hutchings wrote:
> I was replying to Victor. Hopefully it wasn't top-posted...
>
How can you top-post if you don't quote? Arse.

--
Ian Collins

Victor Bazarov

unread,
Sep 15, 2014, 3:20:57 PM9/15/14
to
On 9/15/2014 3:09 PM, Robert Hutchings wrote:
> Yes, I misspelled SIGH. What is the penalty for misspelling in this group?

This newsgroup is unmoderated. There are no penalties beyond a well
deserved place in a killfile. However, if everybody places you in their
killfiles, it essentially means boycott. People will ignore you if you
irritate them too much. I think the times of reporting somebody to
their ISP upon severe violations of newsgroup netiquette has gone the
way of dinosaurs.

Acting like an ass when you're "new people" (from your own admission)
does irritate some folks here, and will likely land you in killfiles.
That usually means you won't be able to get real help if you ever need
it. Of course, c.l.c++ is but one resource among so many on the 'net
that inability to get help from c.l.c++ should not really discourage any
ass out there...

When you come to a newsgroup it is customary to ask for the FAQ. Even
though you didn't ask, I'll suggest http://www.parashift.com/c++-faq/ ,
perhaps it will be of some use to you.

Drew Lawson

unread,
Sep 15, 2014, 3:22:59 PM9/15/14
to
In article <da324135-79e7-4c4e...@googlegroups.com>
Robert Hutchings <rm.hut...@gmail.com> writes:
>I was replying to Victor...OH, RIGHT, I top-posted. My bad.

Um, no, you did not.

If you are going to through a fit about someone's sig quote regarding
top posting, at least bother to learn what the phrase means.

--
Drew Lawson | If dreams were thunder,
| and lightning was desire,
| This old house would have burnt down
| a long time ago
Message has been deleted

Dombo

unread,
Sep 15, 2014, 3:30:44 PM9/15/14
to
Op 15-Sep-14 18:05, Robert Hutchings schreef:
> I will take 100 lashes later today...

Kinky

Dombo

unread,
Sep 15, 2014, 3:34:34 PM9/15/14
to
Op 15-Sep-14 21:09, Robert Hutchings schreef:
> Yes, I misspelled SIGH. What is the penalty for misspelling in this group?

To repent you will have to refrain from masturbating for one week.


Ian Collins

unread,
Sep 15, 2014, 3:36:02 PM9/15/14
to
Robert Hutchings wrote:
> The word you are looking for is "ass" Ian. "Ass".

Ass is a synonym for donkey.

No one uses "Arse" except old British guys who use British slang even
when they know its stupid. Must be a British thing...

It's a not-American thing.

--
Ian Collins

David Brown

unread,
Sep 16, 2014, 4:34:27 AM9/16/14
to
On 15/09/14 18:05, Robert Hutchings wrote:
> Wow. I daresay I won't be posting in this group of super-experts
> again. Please pardon my ignorance and stupidity. I will take 100
> lashes later today...
>

You can take your 100 lashes for posting to a newsgroup without
understanding Usenet or its conventions. Please get yourself a proper
newsreader, and learn to quote correctly and post correctly. You'll
find it is far easier to hold a conversation.

I don't think anyone here has been trying to be rude or unhelpful - but
when you ask a silly question, you get a silly answer. And your
question /was/ silly, as you would have found out with a couple of
minutes googling or looking on Wikipedia. Those posters (including
myself) who have tried to give you a little more information have made
it clear that C++ is not particularly relevant to HFT.

David

David Brown

unread,
Sep 16, 2014, 4:40:50 AM9/16/14
to
Also, "arse" is not slang - it is the correct word. "ass", as used by
Americans, /is/ slang - it does not come from the real English language
word (meaning "donkey"), but is a made-up word to allow pseudo-polite
Americans to swear while pretending that they are not swearing. It is
in the same category as other hypocrisies such as "gosh", "darn", or "oh
sugar".

In this case, however, both "arse" and "ass" (using its real meaning as
a word) are appropriate.

Jorgen Grahn

unread,
Sep 17, 2014, 8:42:53 AM9/17/14
to
On Mon, 2014-09-15, Victor Bazarov wrote:
...
> "Down with newsgroup bullying!!!" -- I can just see CNN spending hours a
> day chewing fat on that... <sigh>

Replace "newsgroup" with "social media", and I'm pretty sure they are
doing just that. Here in .se they have (to confuse things) even
appropriated the term "troll" for those bullies ...

/Jorgen

--
// Jorgen Grahn <grahn@ Oo o. . .
\X/ snipabacken.se> O o .

Paavo Helde

unread,
Sep 20, 2014, 1:58:53 PM9/20/14
to
sc...@slp53.sl.home (Scott Lurndal) wrote in
news:NOERv.218997$No4.1...@fx19.iad:

> Algorithmic processing is often offloaded to custom coprocessors
> (generally Field-Programmable Gate Arrays, which are programmed in
> Verilog/VHDL).
[...]
> Given the low-latency requirements, C++ may not be the best choice
> anyway.

FYI: it appears that there are some systems which claim to be able to
program FPGA directly from a C or C++ program:

http://www.xilinx.com/support/documentation/sw_manuals/ug998-vivado-intro-
fpga-design-hls.pdf

I have not used it (or even read through the pdf) so don't know how much
merit it has.

Cheers
Paavo


Scott Lurndal

unread,
Sep 22, 2014, 10:48:07 AM9/22/14
to
IIRC, the available C++ language features are severely constrained.
0 new messages