10000rpm hard drive?
- Thread starter a529612
- Start date
What I am saying and have been expressing in other threads is that you do not need all the fancy bells and whistles to be profitable. Many traders who are just starting in this field, try to surround themselves with 5k rigs thinking it will give them an advantage. I admit I used to think that I needed the best processor, 4 gig of ram, and a $500 video card to be able to trade. Then I met a guy who is pulling 2k a day trading on a Pentium Celeron M with 512Mb of RAM and a 80 gig 5400 HD. with one 17" monitor. I could have a 10k PC and could never touch his level even if I had 10 30" monitors connected to it. Basically, it is not about the machine it is about your brain.
Now, do I enjoy the latest gadgets? Of course and I have nothing against Alienware rigs, but I think it should be noted that you do NOT need anything fancy to make serious money trading. I think it is important to stress the above, so that when somebody fresh off the street comes lurking on ET that 5k spent on a shiny PC will not buy him a ticket to being profitable.
There are two issues you are looking at here, do not mix them up and confuse the discussion between them:
1. Data read/write speed
2. Total computing performance speed
To see whether a hdd meets your computing needs, you will find the following spec useful:
1. cache size
2. transfer rate
3. latency
4. spindle speed
RPM is only one of many factors that affects your harddisk performance, which might (or might not) impact your total computing performance.
At the heart of a faster RPM harddisk, you are hoping for a faster data read/write access.
Does your disk has a transfer rate that is fast enough to meet your need? You can find out here:
http://www.seagate.com/www/en-us/products/
http://westerndigital.com/en/products/index.asp?Cat=3&Language=en
just click on the data sheet and compare.
Further self searching questions:
1. What is your requirement for data transfer in MB per second? i.e. how many MB does your application writes to the disk per second?
e.g. if your datafeed sends you 10 updates per second, and each update/symbol consists of 100 bytes of data, and you are tracking 100 symbols. (I am way over estimating here.)
10 x 100 x 100 = 100,000 bytes per second.
2. Can the cache/buffer hold your average read/write packet? What does it mean if my answer is "Yes"?
hehehe
You got me Giles117,
I am a system trader with multi entries and exits conditions all program into the system, I do this, because one single simple system won't win all the time, and if you are a turely mechnically trader, then you would have to program everything on you chart.
I reach a point where computer performance is needed, I have a few sets of indictors per entry and some of the codes are complex - such as daily indictors. On my volume and on the market I trade (HSI), 2 sec off a chart would hurt profitability. I had done my time in building my trading system (3 months, full time), risk and money management (last 7 months), and now I am looking into hardwares, coz this the weakest link rightnow.
I am just surpise people piss at the idea of better hardware. Certain amount of slippage over a month will cost you more then 2 10k harddisk. Sounded like some trader here would rather spent more on a set of Mags for their ride instead of trading computer, the machine that acturally help them make a living.
I got aggressive with Saxon22, coz if he has any health respect to the market, he would give it his all, write a complex system that stress his computer to the max, buy the best hardware he can afford, and go for it (ie, trade real money). Instead of giving out smartass reply, sitting there paper trading...
Quote from saxon22:
What I am saying and have been expressing in other threads is that you do not need all the fancy bells and whistles to be profitable. Many traders who are just starting in this field, try to surround themselves with 5k rigs thinking it will give them an advantage. I admit I used to think that I needed the best processor, 4 gig of ram, and a $500 video card to be able to trade. Then I met a guy who is pulling 2k a day trading on a Pentium Celeron M with 512Mb of RAM and a 80 gig 5400 HD. with one 17" monitor. I could have a 10k PC and could never touch his level even if I had 10 30" monitors connected to it. Basically, it is not about the machine it is about your brain.
Now, do I enjoy the latest gadgets? Of course and I have nothing against Alienware rigs, but I think it should be noted that you do NOT need anything fancy to make serious money trading. I think it is important to stress the above, so that when somebody fresh off the street comes lurking on ET that 5k spent on a shiny PC will not buy him a ticket to being profitable.
Sax,
I understand what you say, let's just say if you put enough of your brain into codes, it will stress the charting software and computer hardware. I have outdone Metastock 9.1 to use enough external functions to actually slow the computer down.
If you happens to DIY computers, then go for gamers' hardware such as those massive heatsink and other craps, they are cheap, and why wouldn't you want your CPU to be cooler etc? This is a machine that will help you make a living, why not go overbroad with speed and reliability?
Call me crazy coz I will go for watercool CPU and harddisks for my next workstation.
Sax, just go for it, believe in your system and just go for it! As I said on my PM to you a long time ago, give it your best and just GO FOR IT.
I'm not confusing anything, I'm responding to the discussion in this thread. No one needs a 10k drive for a typical trading PC and anyone that really does need that kind of IO performance should not be consulting an ET thread for advice.
Your suggestion to look at drive manufacturer data sheets for performance stats is a bad idea; the transfer rates listed there don't represent real-world performance. Better to look at benchmarks from a site that does drive testing. StorageReview.com used to be good but they seem to be going downhill.
Most people have no idea what their drive I/O requirements are and your simplistic method of how to compute it falls way short of what is required if someone wanted to undertake such an effort.
I wonder why they haven't invented a cordless USB thumb drive? Without a cord to contend with, that would mean one less moveable/attachable part to mess around with. Imagine plugging a small Corvair thumb drive into a USB port and it just "hangs there" and "sticks out" from the computer by a few inches or whatever. Since it doesn't have a cord, it would "stick out" from the computer at a 90-degree angle, but who cares?
Rhetorical question: Would people rather carry around a thumb drive and a cord in their pockets, or would they prefer to carry just a thumb drive? Hmm, I don't know. Perhaps I should be an inventor instead of a trader, lol.
+-*/ Math_Wiz