Get two free computers for a year!
 

Hey all... I'm not sure if this is a good idea or not. But what the heck...

I'm at the stage with my project where I need some BIG iron. But only occasionally. I decided it was time to start learning how to leverage Amazon's EC2.

I went to sign up, and discovered that for new customers they're offering a year's worth of "t1.micro" Linux instances AND a year's worth of Windows instances for free! All you need is a valid credit card -- they won't bill it unless you have "overage".

Details are at [URL="http://aws.amazon.com/free/faqs/"]http://aws.amazon.com/free/faqs/[/URL].

The virtual machine provided isn't too powerful (and its performance jumps around a bit, as one would expect), but it's definitely not a 386 either... And it's free!

[CODE][Work thread Feb 3 18:50] Timing 44 iterations of 768K FFT length. Best time: 13.440 ms., avg time: 17.524 ms.
[Work thread Feb 3 18:50] Timing 38 iterations of 896K FFT length. Best time: 16.233 ms., avg time: 19.454 ms.
[Work thread Feb 3 18:50] Timing 33 iterations of 1024K FFT length. Best time: 18.026 ms., avg time: 22.605 ms.
[Work thread Feb 3 18:51] Timing 26 iterations of 1280K FFT length. Best time: 28.171 ms., avg time: 31.836 ms.
[Work thread Feb 3 18:51] Timing 22 iterations of 1536K FFT length. Best time: 32.071 ms., avg time: 34.310 ms.
[Work thread Feb 3 18:51] Timing 19 iterations of 1792K FFT length. Best time: 37.485 ms., avg time: 40.224 ms.
[Work thread Feb 3 18:51] Timing 16 iterations of 2048K FFT length. Best time: 43.034 ms., avg time: 44.666 ms.
[Work thread Feb 3 18:51] Timing 13 iterations of 2560K FFT length. Best time: 52.075 ms., avg time: 55.121 ms.
[Work thread Feb 3 18:51] Timing 11 iterations of 3072K FFT length. Best time: 66.438 ms., avg time: 74.521 ms.
[Work thread Feb 3 18:51] Timing 10 iterations of 3584K FFT length. Best time: 92.998 ms., avg time: 108.097 ms.
[Work thread Feb 3 18:51] Timing 10 iterations of 4096K FFT length. Best time: 924.122 ms., avg time: 1065.588 ms.
[Work thread Feb 3 18:52] Timing 10 iterations of 5120K FFT length. Best time: 117.753 ms., avg time: 136.621 ms.
[Work thread Feb 3 18:54] Timing 10 iterations of 6144K FFT length. Best time: 141.004 ms., avg time: 146.417 ms.
[Work thread Feb 3 18:55] Timing 10 iterations of 7168K FFT length. Best time: 162.681 ms., avg time: 256.381 ms.
[Work thread Feb 3 18:56] Timing 10 iterations of 8192K FFT length. Best time: 1858.992 ms., avg time: 1979.887 ms.
[Work thread Feb 3 18:56] Timing trial factoring of M35000011 with 61 bit length factors. Best time: 4.001 ms.
[Work thread Feb 3 18:56] Timing trial factoring of M35000011 with 62 bit length factors. Best time: 4.017 ms.
[Work thread Feb 3 18:56] Timing trial factoring of M35000011 with 63 bit length factors. Best time: 7.024 ms.
[Work thread Feb 3 18:56] Timing trial factoring of M35000011 with 64 bit length factors. Best time: 6.895 ms.
[Work thread Feb 3 18:56] Timing trial factoring of M35000011 with 65 bit length factors. Best time: 6.189 ms.
[Work thread Feb 3 18:56] Timing trial factoring of M35000011 with 66 bit length factors. Best time: 6.210 ms.
[Work thread Feb 3 18:56] Timing trial factoring of M35000011 with 67 bit length factors. Best time: 6.109 ms.
[Work thread Feb 3 18:56] Timing trial factoring of M35000011 with 75 bit length factors. Best time: 6.565 ms.
[Work thread Feb 3 18:56] Timing trial factoring of M35000011 with 76 bit length factors. Best time: 6.493 ms.
[Work thread Feb 3 18:56] Timing trial factoring of M35000011 with 77 bit length factors. Best time: 6.500 ms.
[/CODE]

Certainly enough to clear a few DCs, LLs or P-1s in a year.... :smile:

Very interesting. Thanks!

It's awesome! Thanks for sharing!

[CODE]CPU Information:
Intel(R) Xeon(R) CPU E5430 @ 2.66GHz
CPU speed: 2659.55 MHz
CPU features: Prefetch, SSE, SSE2, SSE4
L1 cache size: 32 KB
L2 cache size: 6 MB
[/CODE]
Almost identical values in benchmark.
Fairly nice (even if a bit old, this series was launched in 2007), and, indeed, free!

We don't look a Trojan horse in the mouth, do we?

750 hours is just a little over a month. Or am I reading it wrong?

But, thanks for sharing in any way. Nice find.

[QUOTE=TObject;328197]750 hours is just a little over a month. Or am I reading it wrong?][/QUOTE]

24 * 31 == 744.

[QUOTE=chalsall;328198]24 * 31 == 744.[/QUOTE]

I mean, it is not anywhere close to a year; only 750 hours are free. Or not?

[QUOTE=TObject;328199]I mean, it is not anywhere close to a year; only 750 hours are free. Or not?[/QUOTE]

You didn't read the Terms...

You get 750 hours per [B][I][U]month[/U][/I][/B] for a [B][I][U]year[/U][/I][/B].

Sadly, it appears English is as scary as Math.

Just introduce Latin, and the masses will be running scared towards to the Super Bowl....

[QUOTE=chalsall;328200]You didn't read the Terms...

You get 750 hours per [B][I][U]month[/U][/I][/B] for a [B][I][U]year[/U][/I][/B].

Sadly, it appears English is as scary as Math.

Just introduce Latin, and the masses will be running scared towards to the Super Bowl....[/QUOTE]

Thank you for answering my question. The terms, as posted on the amazon web site, were unclear to me.

:max:

Now the physics question: how fast does the server need to move relative to the observer so that the when the observer counts 744 hours on his or her watch, 750 hours had passed on the server. LOL

[QUOTE=TObject;328203]Now the physics question: how fast does the server need to move relative to the observer so that the when the observer counts 744 hours on his or her watch, 750 hours had passed on the server. LOL[/QUOTE]

A bullshit question.

Given that the server and the user are (mostly) in the same time-space continuum, 744 hours of the user's time will be *very* close to the server's observed time.

[QUOTE=chalsall;328204]A bullshit question.[/QUOTE]

You are no fun.

[QUOTE=TObject;328205]You are no fun.[/QUOTE]

Agreed.

I can't visualize how this service would work. Would you sign into something and your monitor and keyboard would appear to be hooked to an Amazon computer?

It could be as easy as that.
They even have a Java-based MindTerm that let's you ssh.

Windows session (that I expected to be even easier) didn't let me remote-login yet. Will debug later (from home); maybe it's our institute's firewall.

[QUOTE=Chuck;328213]I can't visualize how this service would work. Would you sign into something and your monitor and keyboard would appear to be hooked to an Amazon computer?[/QUOTE]

Very far from.

Absolutely no disrespect intended, but unless you know how to use SSH don't bother attempting.

Chris, granted there's documentation there (even too much), but there are some silly traps set up everywhere in their interface.

Would you write a short and sweet step-by-step tutorial for the rest of us?

I've stumbled through some ditches (and will try to learn better later), but here are a couple things that I would have already done differently if I did everything from the start:
1. The default location of Virginia escaped my attention. It is only on the second try that I switched a drop-down box in the wizard to north California.
2. I should have probably walked though IAM wizard first. Because I didn't, the role on the wizard log pages defaults to "None"
3. Do you have recommendations on which tiny instance to choose?
4. Can you give an advice how to make jobs nohup if the terminal goes away?

[COLOR=green]Updates: [/COLOR]
[COLOR=green]1. nohup works[/COLOR]
[COLOR=green]2. imported ppk into putty[/COLOR]
[COLOR=green]3. ubuntu doesn't have gcc! but it is relatively easily installed. works ok[/COLOR]

[QUOTE=chalsall;328217]Very far from.

Absolutely no disrespect intended, but unless you know how to use SSH don't bother attempting.[/QUOTE]

I can SSH into the NAS server I have on my network (mildly hacked Buffalo Terastation), but that is Linux; how would you use a Windows instance?

You don't mean a command line Windows interface do you?

[QUOTE=chalsall;328204]A bullshit question.

Given that the server and the user are (mostly) in the same time-space continuum, 744 hours of the user's time will be *very* close to the server's observed time.[/QUOTE]

-100 points for using "time-space continuum". His question was perfectly well defined.

[QUOTE=Chuck;328220]I can SSH into the NAS server I have on my network (mildly hacked Buffalo Terastation), but that is Linux; how would you use a Windows instance?

You don't mean a command line Windows interface do you?[/QUOTE]

Yich. I might have considered trying if it was at least a Linux interface. Good luck with getting it to work :smile:

[QUOTE=Batalov;328219]Chris, granted there's documentation there (even too much), but there are some silly traps set up everywhere in their interface.

Would you write a short and sweet step-by-step tutorial for the rest of us?

I've stumbled through some ditches (and will try to learn better later), but here are a couple things that I would have already done differently if I did everything from the start:
1. The default location of Virginia escaped my attention. It is only on the second try that I switched a drop-down box in the wizard to north California.
2. I should have probably walked though IAM wizard first. Because I didn't, the role on the wizard log pages defaults to "None"
3. Do you have recommendations on which tiny instance to choose?
4. Can you give an advice how to make jobs nohup if the terminal goes away?

[COLOR=green]Updates: [/COLOR]
[COLOR=green]1. nohup works[/COLOR]
[COLOR=green]2. imported ppk into putty[/COLOR]
[COLOR=green]3. ubuntu doesn't have gcc! but it is relatively easily installed. works ok[/COLOR][/QUOTE]

I dug into the pages. It seems that we Europeans have some small clauses to watch against...

Luigi

[QUOTE=ET_;328264]I dug into the pages. It seems that we Europeans have some small clauses to watch against...[/QUOTE]

Such as?

[QUOTE=Batalov;328219]Would you write a short and sweet step-by-step tutorial for the rest of us?[/QUOTE]

OK. Log into your EC2 interface. The default environments are WinBlows or Ubuntu Linux, but CentOS et al are also available.

Make sure your firewall settings allow port 22 to be open. It doesn't appear to be by default. Then create a key pair, and launch the instance.

Next, log in, and run the OS's update.

Then create a user for yourself. (Careful if you're running Linux -- regular user logins might not be enable.) And then load any software you might regularly use.

Then create a new AMI based on this instance. Be aware the virtual machine being imaged will be shut down during this process unless explicitly being told not to.

For the advanced users, a EBS volume might be created which can be mounted under an EC2 instance. Perhaps under /home/.

Both in win and linux instances, I've observed over a day of running that though DC processes can (and occasionally run) at, say, 0.029 s/iter, 99% of the time they run ~10x slower (wobbling between ~0.2 and 0.4 s/iter). At this rate even the smallest DC jobs at ~29M will take three months per job.

I am not sure why the program (or the host's OS) does that. I have thrown the niceness parameter into "10" (sadly, there wasn't an "11" setting). We'll need a non-apologetic version of P95 compiled - i.e. the version that doesn't renice and doesn't throttle on other processes (from other invisible users on the same physical node).

[QUOTE=Batalov;328377]I am not sure why the program (or the host's OS) does that.[/QUOTE]

What part of "virtual machine" isn't clear? :smile:

Well, I would imagine that they oversubscribe the physical nodes like the airlines do. In 2015, we will be lucky to get an upper ticket in a doubled-up seat on any airline. :-) Or lower position on the seat, if you don't mind the weight or rather, prefer a little weight vs something pressing against your lower back. There will be a surcharge for the non-default preference.

Anyway, that's what I meant by non-apologetic. mprime/Prime95 tells every process "Oh! Terribly sorry, Sir, I will wait. No, no, after you! I insist! This virtual node has no real load; please let everything else run first."

Amazon's t1.micro instances allow for short bursts of high CPU usage but anything longer is severely throttled.

They work passably well for hosting basic web sites and similar tasks but even simple operations like launching a decent text editor can trigger throttling.

For Windows instances, running Windows Update always triggers throttling turning it into an even more painfully long effort than usual.

I used a couple of t1.small instances for my business and they are much more usable but of course they come at a cost.

Don Leclair

[QUOTE=dleclair;328419]Amazon's t1.micro instances allow for short bursts of high CPU usage but anything longer is severely throttled.[/QUOTE]

I'm finding them very handy for mounting EBS volumes into which I rsync several thousand images for processing in larger virtual machines....

[QUOTE=chalsall;328372]Such as?[/QUOTE]

Higher costs and lower CPU usages.

I read the other posts, and I decided not to try: the virtual environment doesn't match fine with high requests.

I've been through this when ISPs offered free (but painfully slow) connections.

Luigi

Well, this idea works, even if the ETAs are wrong.
Let's see how good these t1.micros are after 3 months.

So I tried this out to get some experience with EC2, and I have not been impressed with the results. A first time LL on a 62m exponent was not on track to finish within the year, with iteration times of ~.400 seconds each. So I switched to P-1 and I got a "broken pipe" error when trying to enter stage 2. Trial factoring is possible, but is essentially useless considering it is about 100x slower than a GPU.

I set it to try a Double-Check, maybe that will complete within a year...

BTW this was on a Linux micro, the windows one was much much slower and still hasn't finished stage 1 of P-1.

So essentially I feel as if logging into the instances to check them costs more ghz/hr to my desktop than I gain by having the instances.

Meanwhile my two instances didn't die yet. And I didn't visit them for two months, no sirrie.

[QUOTE=Batalov;335439]Meanwhile my two instances didn't die yet. And I didn't visit them for two months, no sirrie.[/QUOTE]

Yeah I am thinking double checks might be the best use for these. I guess you could always transfer a partially completed first time LL...

Is there an idiot's guide (for folks like me) to using these for GIMPS work? I could always use more power.

[QUOTE=Uncwilly;335575]Is there an idiot's guide (for folks like me) to using these for GIMPS work? I could always use more power.[/QUOTE]

It was a fairly easy process, and I am by no means amazing with computers. Basically you go to the ec2 site, make an account and open a micro instance. They have some startup guides there that are fairly straightforward too. You download a pem key and use windows remote desktop to login to the windows one, and an ssh command to login to the Linux one.

[QUOTE=Batalov;335439]Meanwhile my two instances didn't die yet. And I didn't visit them for two months, no sirrie.[/QUOTE]

It appears these finished successfully.
Did you grab two more?

Yep, two more are reaching the end. Every time I remember about them and give a quick look at them, their speed makes me think of
track #2 (Grave Disperato /For wind orchestra and clock/) from [URL="http://en.wikipedia.org/wiki/Tales_and_Songs_from_Weddings_and_Funerals"]Tales and Songs from Weddings and Funerals[/URL].

[YOUTUBE]kCzMZ-IuKTQ[/YOUTUBE]

[QUOTE=TObject;328203]Now the physics question: how fast does the server need to move relative to the observer so that the when the observer counts 744 hours on his or her watch, 750 hours had passed on the server. LOL[/QUOTE]

Requires the server to travel faster than the speed of light. Using currently accepted relativity formulas, there would be no time dilation at warp speed so this is not possible (although there are conflicting theories about time dilation at faster than light travel).

[QUOTE=flashjh;347288]Requires the server to travel faster than the speed of light. (SNIP).[/QUOTE]

Wouldn't it be the other way around--that less time would pass for the server than for the "stationary" observer? (EDIT: .....if the server were the one at relativistic speed.....)

[QUOTE]There was a young lady named Bright
Whose speed was much faster than light
She arose one day in a relative way
And returned on the previous night.[/QUOTE]