I built a bittorent server using an old pentium III M chip. It uses 17 watts AC at idle, including an external 2.5 inch hard drive. The upside is it is very quiet, low power, and can run any distro. The only downside is it doesn't have ECC memory, but as far as I know, no laptop does, and neither do any very low power computer. It has a dvd drive, and thanks to a notebook card, USB2. It doesn't have gigabit ethernet, but it supports a standard PS/2 keyboard, mouse, and VGA external video, so I can use a KVM switch to control it.
You need to name the retailer. Some retailers are great, and some suck. I have reviewed many, explaining the details of my dealings with them.
You also need to name the NAS and what is wrong with it. The way the internet works, everyone gets to benefit from the experiences of others. Saying some nameless NAS sucks doesn't help anyone else. Saying XYZ NAS sucks because of a, b, and c means something.
On my personal web site, I have details of my dealings with many computer hardware and software vendors. Some are really good, and some are really sucko. I had some serious problems with Netgear, and someone else read about it and asked Netgear why their support was so bad, and not getting a decent answer, ended up getting $10k of gear from a different vendor.
Yeah, 600mg of ibuprofen, which is like 3 200mg ibuprofen pills. It is a crazy drug, which will help with headaches and muscle pain. I have never heard of anyone abusing ibuprofen. You can go to costco or the like and get 1000 pills (200mg each) for well under $10.
I personally think it silly to have 'prescription strength' ibuprofen, which is the same as 3 over the counter pills.
Lets say the 13 year old did have some ibuprofen. No abuse potential. No danger to others. Less dangerous than a paperclip or a pencil. No objective reason to strip search, other than a
pseudo-war on prescription drugs.
The school ban on prescription drugs is foolish. The strip search is clueless as best and likely felonious. You would hope the administrators had a clue...
Until the i7 has ecc, there in no winning against the phenom II which has ecc. If you are doing anything excepting gaming, you likely need ecc. Since the i7 doesn't have it (until the i7 based xeons come out), there is simply no comparison.
Like bring a knife to a gunfight, you are fscked.
I am not sure what you are doing with your cluster, but I am sure your cluster is now unreliable. It is most likely your old slow xeons had ECC memory. They might even have had chipkill like all the opterons and phenom II's have. But without ECC,
I suspect you have a very big, somewhat hot computer that has a bunch of random errors.
Making a bunch of assumptions, 20 xeons, 4gb of ram each, means 80gb of memory. Assuming you still are using 80gb of memory, and with IBM's estimate for memory errors (1gb = 1 error per week), you will see have about 80/7, or 11 errors per day. That is almost an error every 2 hours.
There was once a 'cluster' made of 1u macs. Apple ended up replacing every one with a newer model with ecc. It is true that there are fault tolerant algorithms that can deal with low error rates, but if aren't coding with them, I would have little confidence in the correctness of your output.
Perhaps the i7 is cool for a gamer, but when I paid my money, I bought a phenomII 940, because it has chipkill ecc and background memory scrubbing (it scrubs all memory every 8 hours).
This is pretty much mandatory for reliable computing, let alone a cluster.
I did pay full price for my mchale pack, but the rest of the packs I got used, and for a price between a good deal, and a stupid great deal.
I did get some fanny packs custom made of 100% spectra from offshore. They were only $40 each (plus expensive shipping). I am confident they will last forever.
I like gear that lasts for a long time, and I hate to have to send back broken gear for getting fixed. It looks like Tom's bags are very high quality, and should last for many years, unlike my cheap a#@ dell backpack which is wearing out in several places after two years of use.
I understand you using the shorthand. However, it can be a bit misleading as pure dyneema is much more expensive, and to some much more desirable. I noticed that is was often misspelled on your website as dyeema or some such.
I really prefer to call it dyneema ripstop or spectra ripstop or whatever the fabric manufacturer calls it. I happen to like pure spectra backpacks (which are quite expensive and rare), and when I use a search engine looking for 'spectra backpack', I don't want to see the dyneema ripstop stuff. I have reviews of six of my spectra packs at http://www.weasel.com/gear_pack.html
I happen to have a The North Face expensive pack made of spectra ripstop. After two or three uses, it developed two holes in the spectrum fabric, and I returned it for repair. They wouldn't fix it, either under warranty, or with me paying for it. I was a bit upset at that.
I do think it a somewhat odd choice of material for inside a bag. Overall, your bags look to be quite well engineered.
Real Dyneema is ultra high molecular weight polyethylene, and is always white, though it can be waterproofed and dyed to a light grey. It is much more expensive than nylon, and its main claim to fame is it is roughly 10 times more abrasion resistant than nylon. It is commonly used in bullet-proof vests, replacing kevlar as the material of choice.
Since the Tom Bihn site is down, I can't check their use of spectra. A common use is as a ripstop which is used along with nylon. A common name for such ripstop is Spectrum. It is very expensive to coat (as most things don't stick to polyethylene) and always has white grids.
I am not sure what the reviewer is referring to when they say 'bright yellow ("Solar") version of Dyneema that is used for some Bihn products (or in red), but a white-grid-on-grey isn't bad.', but they are clearly not referring to Dyneema fabric, as there is no such thing as yellow or red Dyneema fabric.
the dawn of object oriented programming
on
C# In-Depth
·
· Score: 1
In the fine article, it says:
So C++ I think is a fantastic example of a very appropriate evolution of an existing language. It came right at the dawn of object oriented programming...
Well, lets see, according to wikipedia C++ appeared in 1983. Smalltalk started development in 1969 and was publically available in 1980. And simula appeared in 1967.
So I guess that means that 1983 is the dawn of 1967... I think this guy forgot what object oriented means, and has drank too much C++ coolaid.
The fine book said it was hard to test procedural code. I was replying that that was false.
Clearly you can't test for every possible input, but that isn't what I was addressing. Nor did I ever claim that you could test for every possible input, nor the desirability of doing so.
If I did have a function that took in two doubles, I would try some typical double values, 0, and some corner cases. Only a mo-ron would think to test for every value.
If I was doing white box testing, I would carefully look over the algorithm used, and try to find values to exercise all the conditions of the code.
Though it is hard to test spaghetti code, it is hard no matter if the code is procedural or object-oriented.
If you have reasonably well written procedural code it is easy to unit test. For C, there is http://cutest.sourceforge.net/ a very simple framework. Lets say you just wrote super_sort. You can write tests to test it with random input, sorted input, reverse sorted input, big input, small input, and the like. If you ever break super_sort, you will find out as soon as you
run your unit tests.
IMHO, linear algebra may be complex, but there is generally no state. Purely procedural code is fine for it (though C may not be). There are many simple as well as complex things that have no internal state, and will not benefit from object orientation.
I gave my wife a 4 ring gold puzzle ring. I bought myself a 6 ring silver puzzle ring. They are geeky, and reasonably inexpensive. A silver ring will run $30 - $300 depending on style and thickness. Silver kills bacteria on contact, and can be cut off if needed. The 5 ring puzzle ring is supposed to be harder to solve...
The bottom line, was she liked it, so clearly I made the correct choice.
There are several exceptions in the USA. For example, there is also the http://en.wikipedia.org/wiki/Intermountain serving los angeles. It is a 785 kilometer HV DC line.
I suspect there are move HV DC lines in the US.
However, listing two counterexamples is more than sufficient to counter 'there are none'.
If you read your wikipedia link, you will see a mention and link to the pacific intertie at
http://en.wikipedia.org/wiki/Pacific_DC_Intertie
It is more formally known as the Pacific DC Intertie, and it happend to be a DV long distance transmission line, which you claim doesn't exist.
Perhaps you should read your links...
Here is the link I was looking for yesterday:
http://www.ueverest.com/camp_life.aspx?postId=19
Here is the important part:
The air is very thin and machinery such as hard drives have a very high failure rate in this rarefied atmosphere. Delicate equipment would have to be transported over miles of rugged terrain by Yak or porters. We are very much isolated from the usual consumer grid, so all the equipment we need would have to last for the entire trip or be repaired in the field and there would be no chance to purchase or replace broken hardware in this remote location.
The Hardware
Laptop computers provide most of the computer power we need on the mountain but they have to stay lower on the mountain, at or below advanced base camp at 21,000 feet. For redundancy we brought three laptops, all different brands, just in case one particular brand proved problematic. The primary mode of failure for laptops in this environment is hard drive failure since hard drives rely upon the viscosity of air to provide lubrication and damping among the moving parts, the same manner that oil provides lubrication and damping for moving machine or engine parts.
Due to low pressure and cold temperatures, the viscosity of air is a fraction of what it is at sea level and hard drive failure is common. Each laptop had two spare hard drives, each hard drive had the operating system, and all of the mission critical software pre-installed upon it before we left. When hard drives fail we could change out a new one in several minutes since they were pre-built. Data was backed up daily to either CD/DVD or to flash drives, since these media were much less subject to failure at altitude. It didnt take long for the first hard drive to fail - we lost the first hard drive at a mere 12,500 feet en-route to base camp.
You will want to use a solid state disk when you are at Everest base camp. If you read about computers used there, the hard drives fail very quickly due to low air pressure. Hard drives are not rated to work at 18,000 feet.
You can buy an AC Propulsion eBox (converted Scion xB) for $55k plus the Scion cost. Ship it to europe. Much cheaper than the
tesla at $150k. the eBox does 0-60 in 7 seconds, so it isn't a slug, though not as fast as a Tesla.
And it seats more than 2 people.
I have sent in several hundred rebates. A few (mouse systems) were not honored. The retailer (CompUSA) paid me for those rebates, and told me they stopped carrying mouse systems products. Then I bought two items from Vastec at fry's. I got neither rebate back. I emailed the company. They ignored me. I called the company (nobody ever answered). This is not an isolated incident. This is corporate policy. I ended up giving copies of the rebates to fry's. They sent them to fry's corporate, and they say I will get my money back. I noted this at http://weasel.com/comp-rebate.html#vastech . By the way, Microsoft always pays their rebates, and does so quickly. I have the results of my rebates by company at http://weasel.com/comp-rebate.html
I was also 'lucky' enough to play around with a Niagara processor. I happened to have my own highly threaded app. It ran really, really slowly. The sun engineers eventually figured out that two of my threads would saturate one of their processor cores. (there was zero fp in my code).
With 32 threads, their fastest T2000 took 10,600 seconds. My old, dual xeon 2.4ghz machine took 11,494 seconds. An 8 way opteron 2.8ghz machine was 11.6 times faster than the niagara. A a Sun Sparc V890 8cpu, 16 core was8.3 times faster than the niagara.
Since their t2000 cost around $20k, and a dual 2.4ghz xeon cost around $500-$1000, that made the price/performance roughly 20-40 times better for a xeon than the niagara.
Perhaps their new chip is higher performing than their earlier generation.
You can see all the gory details at www.weasel.com/comp-perf.html source code available upon request.
Slashdot Mirror
Actually, if we would have funded Project Orion, we would have gone to Saturn in the early 1960's. See http://www.ted.com/talks/george_dyson_on_project_orion.html among other references.
I built a bittorent server using an old pentium III M chip. It uses 17 watts AC at idle, including an external 2.5 inch hard drive. The upside is it is very quiet, low power, and can run any distro. The only downside is it doesn't have ECC memory, but as far as I know, no laptop does, and neither do any very low power computer. It has a dvd drive, and thanks to a notebook card, USB2. It doesn't have gigabit ethernet, but it supports a standard PS/2 keyboard, mouse, and VGA external video, so I can use a KVM switch to control it.
You can read about my quest for a low power bittorrent server at http://www.weasel.com/comp_server.html#bit2
You need to name the retailer. Some retailers are great, and some suck. I have reviewed many, explaining the details of my dealings with them.
You also need to name the NAS and what is wrong with it. The way the internet works, everyone gets to benefit from the experiences of others. Saying some nameless NAS sucks doesn't help anyone else. Saying XYZ NAS sucks because of a, b, and c means something.
On my personal web site, I have details of my dealings with many computer hardware and software vendors. Some are really good, and some are really sucko. I had some serious problems with Netgear, and someone else read about it and asked Netgear why their support was so bad, and not getting a decent answer, ended up getting $10k of gear from a different vendor.
If you want to see all my computer related reviews and comments, check out http://www.weasel.com/comp.html
Woo hoo!
I didn't know seagate had any 1gb usb freeagent disks. Or perhaps you mean 1tb...
The size limit is 4gb, not 6gb...
the first cray 1 didn't have ecc. the mtbf was 8 hours.
all crays since that have had ecc. seymore cray was a smart dude.
Yeah, 600mg of ibuprofen, which is like 3 200mg ibuprofen pills. It is a crazy drug, which will help with headaches and muscle pain. I have never heard of anyone abusing ibuprofen. You can go to costco or the like and get 1000 pills (200mg each) for well under $10.
I personally think it silly to have 'prescription strength' ibuprofen, which is the same as 3 over the counter pills.
Lets say the 13 year old did have some ibuprofen. No abuse potential. No danger to others. Less dangerous than a paperclip or a pencil. No objective reason to strip search, other than a pseudo-war on prescription drugs.
The school ban on prescription drugs is foolish. The strip search is clueless as best and likely felonious. You would hope the administrators had a clue...
Until the i7 has ecc, there in no winning against the phenom II which has ecc. If you are doing anything excepting gaming, you likely need ecc. Since the i7 doesn't have it (until the i7 based xeons come out), there is simply no comparison. Like bring a knife to a gunfight, you are fscked.
I am not sure what you are doing with your cluster, but I am sure your cluster is now unreliable. It is most likely your old slow xeons had ECC memory. They might even have had chipkill like all the opterons and phenom II's have. But without ECC, I suspect you have a very big, somewhat hot computer that has a bunch of random errors.
Making a bunch of assumptions, 20 xeons, 4gb of ram each, means 80gb of memory. Assuming you still are using 80gb of memory, and with IBM's estimate for memory errors (1gb = 1 error per week), you will see have about 80/7, or 11 errors per day. That is almost an error every 2 hours.
There was once a 'cluster' made of 1u macs. Apple ended up replacing every one with a newer model with ecc. It is true that there are fault tolerant algorithms that can deal with low error rates, but if aren't coding with them, I would have little confidence in the correctness of your output.
Perhaps the i7 is cool for a gamer, but when I paid my money, I bought a phenomII 940, because it has chipkill ecc and background memory scrubbing (it scrubs all memory every 8 hours). This is pretty much mandatory for reliable computing, let alone a cluster.
I did pay full price for my mchale pack, but the rest of the packs I got used, and for a price between a good deal, and a stupid great deal.
I did get some fanny packs custom made of 100% spectra from offshore. They were only $40 each (plus expensive shipping). I am confident they will last forever.
I like gear that lasts for a long time, and I hate to have to send back broken gear for getting fixed. It looks like Tom's bags are very high quality, and should last for many years, unlike my cheap a#@ dell backpack which is wearing out in several places after two years of use.
I understand you using the shorthand. However, it can be a bit misleading as pure dyneema is much more expensive, and to some much more desirable. I noticed that is was often misspelled on your website as dyeema or some such.
I really prefer to call it dyneema ripstop or spectra ripstop or whatever the fabric manufacturer calls it. I happen to like pure spectra backpacks (which are quite expensive and rare), and when I use a search engine looking for 'spectra backpack', I don't want to see the dyneema ripstop stuff. I have reviews of six of my spectra packs at http://www.weasel.com/gear_pack.html
I happen to have a The North Face expensive pack made of spectra ripstop. After two or three uses, it developed two holes in the spectrum fabric, and I returned it for repair. They wouldn't fix it, either under warranty, or with me paying for it. I was a bit upset at that.
I do think it a somewhat odd choice of material for inside a bag. Overall, your bags look to be quite well engineered.
Buy a promise tx2-100 or tx2-133 pci card for it. They are $10 - $15 from ebay, and will allow you to use drives bigger than 137gb.
Of course, you can use scsi drives as big as you want, or even (shudder) a scsi to ide converter to get your big disk lovin'.
Real Dyneema is ultra high molecular weight polyethylene, and is always white, though it can be waterproofed and dyed to a light grey. It is much more expensive than nylon, and its main claim to fame is it is roughly 10 times more abrasion resistant than nylon. It is commonly used in bullet-proof vests, replacing kevlar as the material of choice.
Since the Tom Bihn site is down, I can't check their use of spectra. A common use is as a ripstop which is used along with nylon. A common name for such ripstop is Spectrum. It is very expensive to coat (as most things don't stick to polyethylene) and always has white grids.
I am not sure what the reviewer is referring to when they say 'bright yellow ("Solar") version of Dyneema that is used for some Bihn products (or in red), but a white-grid-on-grey isn't bad.', but they are clearly not referring to Dyneema fabric, as there is no such thing as yellow or red Dyneema fabric.
For info referring to backpacks made of spectra, see http://www.weasel.com/gear_kelty_spectra.html or http://www.weasel.com/gear_pack_mchale.html
In the fine article, it says:
So C++ I think is a fantastic example of a very appropriate evolution of an existing language. It came right at the dawn of object oriented programming...
Well, lets see, according to wikipedia C++ appeared in 1983. Smalltalk started development in 1969 and was publically available in 1980. And simula appeared in 1967.
So I guess that means that 1983 is the dawn of 1967... I think this guy forgot what object oriented means, and has drank too much C++ coolaid.
The fine book said it was hard to test procedural code. I was replying that that was false. Clearly you can't test for every possible input, but that isn't what I was addressing. Nor did I ever claim that you could test for every possible input, nor the desirability of doing so. If I did have a function that took in two doubles, I would try some typical double values, 0, and some corner cases. Only a mo-ron would think to test for every value. If I was doing white box testing, I would carefully look over the algorithm used, and try to find values to exercise all the conditions of the code.
Though it is hard to test spaghetti code, it is hard no matter if the code is procedural or object-oriented. If you have reasonably well written procedural code it is easy to unit test. For C, there is http://cutest.sourceforge.net/ a very simple framework. Lets say you just wrote super_sort. You can write tests to test it with random input, sorted input, reverse sorted input, big input, small input, and the like. If you ever break super_sort, you will find out as soon as you run your unit tests. IMHO, linear algebra may be complex, but there is generally no state. Purely procedural code is fine for it (though C may not be). There are many simple as well as complex things that have no internal state, and will not benefit from object orientation.
I gave my wife a 4 ring gold puzzle ring. I bought myself a 6 ring silver puzzle ring. They are geeky, and reasonably inexpensive. A silver ring will run $30 - $300 depending on style and thickness. Silver kills bacteria on contact, and can be cut off if needed. The 5 ring puzzle ring is supposed to be harder to solve... The bottom line, was she liked it, so clearly I made the correct choice.
There are several exceptions in the USA. For example, there is also the http://en.wikipedia.org/wiki/Intermountain serving los angeles. It is a 785 kilometer HV DC line.
I suspect there are move HV DC lines in the US. However, listing two counterexamples is more than sufficient to counter 'there are none'.
If you read your wikipedia link, you will see a mention and link to the pacific intertie at http://en.wikipedia.org/wiki/Pacific_DC_Intertie It is more formally known as the Pacific DC Intertie, and it happend to be a DV long distance transmission line, which you claim doesn't exist. Perhaps you should read your links...
Perhaps a factor of 10^6, like the difference between 1 cubic meter and 1 cubic centimeter. Do the math.
Here is the link I was looking for yesterday: http://www.ueverest.com/camp_life.aspx?postId=19 Here is the important part: The air is very thin and machinery such as hard drives have a very high failure rate in this rarefied atmosphere. Delicate equipment would have to be transported over miles of rugged terrain by Yak or porters. We are very much isolated from the usual consumer grid, so all the equipment we need would have to last for the entire trip or be repaired in the field and there would be no chance to purchase or replace broken hardware in this remote location. The Hardware Laptop computers provide most of the computer power we need on the mountain but they have to stay lower on the mountain, at or below advanced base camp at 21,000 feet. For redundancy we brought three laptops, all different brands, just in case one particular brand proved problematic. The primary mode of failure for laptops in this environment is hard drive failure since hard drives rely upon the viscosity of air to provide lubrication and damping among the moving parts, the same manner that oil provides lubrication and damping for moving machine or engine parts. Due to low pressure and cold temperatures, the viscosity of air is a fraction of what it is at sea level and hard drive failure is common. Each laptop had two spare hard drives, each hard drive had the operating system, and all of the mission critical software pre-installed upon it before we left. When hard drives fail we could change out a new one in several minutes since they were pre-built. Data was backed up daily to either CD/DVD or to flash drives, since these media were much less subject to failure at altitude. It didnt take long for the first hard drive to fail - we lost the first hard drive at a mere 12,500 feet en-route to base camp.
You will want to use a solid state disk when you are at Everest base camp.
If you read about computers used there, the hard drives fail very quickly due to low air pressure.
Hard drives are not rated to work at 18,000 feet.
You can buy an AC Propulsion eBox (converted Scion xB) for $55k plus the Scion cost. Ship it to europe. Much cheaper than the tesla at $150k. the eBox does 0-60 in 7 seconds, so it isn't a slug, though not as fast as a Tesla. And it seats more than 2 people.
I have sent in several hundred rebates. A few (mouse systems) were not honored. The retailer (CompUSA) paid me for those rebates, and told me they stopped carrying mouse systems products. Then I bought two items from Vastec at fry's. I got neither rebate back. I emailed the company. They ignored me. I called the company (nobody ever answered). This is not an isolated incident. This is corporate policy. I ended up giving copies of the rebates to fry's. They sent them to fry's corporate, and they say I will get my money back. I noted this at http://weasel.com/comp-rebate.html#vastech . By the way, Microsoft always pays their rebates, and does so quickly. I have the results of my rebates by company at http://weasel.com/comp-rebate.html
I was also 'lucky' enough to play around with a Niagara processor.
I happened to have my own highly threaded app.
It ran really, really slowly. The sun engineers eventually figured out that two of my
threads would saturate one of their processor cores.
(there was zero fp in my code).
With 32 threads, their fastest T2000 took 10,600 seconds.
My old, dual xeon 2.4ghz machine took 11,494 seconds.
An 8 way opteron 2.8ghz machine was 11.6 times faster than the niagara.
A a Sun Sparc V890 8cpu, 16 core was8.3 times faster than the niagara.
Since their t2000 cost around $20k, and a dual 2.4ghz xeon cost around $500-$1000,
that made the price/performance roughly 20-40 times better for a xeon than the niagara.
Perhaps their new chip is higher performing than their earlier generation.
You can see all the gory details at www.weasel.com/comp-perf.html
source code available upon request.