>>>>>
<< While Modus claims to be able to divine value for everyone (no matter what they do with their PC), I take the approach that it depends what you do with your PC >>
Does anyone really need SCSI?
<<<<<
Does anyone really need a drive faster than 5400RPM? Does anyone really need a modern PC when a cheap P133 will allow them to run office software, while a gaming console will allow them to play games?
>>>>>
All I'm saying is that a person who bought a full SCSI setup is either not concerned with practical price/performance, or they made a mistake. Surely you can agree with this.
<<<<<
I agree that people who went SCSI were not concerned with YOUR price/performance model. But I do not believe that this is the ultimate mode. I think it's fine if you want to use it, but don't try to convince us to use it too.
>>>>>
You SCSI fanatics, to a person, readily admit that the main attraction of the technology in your eyes was not its practicality or its logical value, but its performance, and specifically, how you felt when you experienced that performance. Yet you still quibble when I claim you weren't making a rational price/performance evaluation?
<<<<<
I had not used a SCSI sytem before I set up mine, so this gushy feeling you keep bringing up could not have been an influencing factor.
>>>>>
<< Who is this "we" anyway? >>
The vast majority of sensible consumers who find it frivolous to pay an extra 300% for an extra 46% of unnecessary performance.
<<<<<
I thought you had no idea how many people follow your price/performance mode.
>>>>>
<< Note that this is a recursive definition. Also note that you are using only the 4th definition given >>
The other definitions of value basically fall into two categories: monetary or practical worth, and morals. Neither of those apply to our discussion, so it's obvious I wasn't using value in that sense. And the dictionary's definition of value is not recursive. It's simply building on an alternate definition. The meaning is quite clear: Equivalent return for value received. If you have trouble with that, Fredrick, then by all means take it up with the Editorial Advisory Committee of the American College Dictionary, Randome House, New York. Fair enough?
<<<<<
If that definition refers to the other definitions, shouldn't they all be supplied? That was my original point. But ok, fine, let's say this definition supports your ideas about value. Another definition has already been supplied in this thread which supports OUR ideas about value.
>>>>>
On the contrary, it is the key to sensible spending toward a practical purporse. It's certainly more meaningful that yours (Value = Enjoyment, Period.)
<<<<<
I have not supplied any equations.
>>>>>
<< I see nothing demented or wrong with purchasing the best you can afford >>
Then we should stop this right now because it's clear you don't have the first clue how a person manages their funds most efficiently.
<<<<<
what good is it to forever manage your funds "most efficiently" by your definition of efficiency which dictates that you never purchase more than you need, or anything without a linear price/performance comparison, if you can never spend your well managed funds on anything better?
>>>>>
<< >>>>> . . . The value purchaser will stop there and buy the product right at the foot of this curve, saving the money for something else with a higher rate of return.<<<<< What if there's nothing left?
Don't be an idiot. There's always something useful left to do you with your money.
<<<<<
Of course. But say you want to upgrade your computer. You already have every component at the top of its price/performance curve, and you're still not satisfied with your performance, do you just fail to upgrade and hope the good stuff gets cheaper eventually?
>>>>
<< we have simply stated that SCSI was worth it for us >>
Only because you were dead set on top PC performance at any cost. That's not sensible or practical: that's a hobby. The problem is, you present it as a triumph.
<<<<<
Not at any cost, just top PC performance within our computing budget.
>>>>>
<< I guess you haven't checked the latest HDtach "poll" on anandtech. I quickly went through it. . . Didn't see a single IDE drive with .6% cpu utilization. >>
You know full well where I got that number. It's straight from any StorageReview drive article. If you want to call them liars, that's fine with me. But here's a recent SR "Elsewhere" quote just to bury your claim that CPU usage is still an issue:
". . . the claim that "IDE is using [sic] the system processor for most actions" is plainly false. These beliefs are merely holdovers from the old days where ATA's DMA capabilities weren't as well developed as that of SCSI."
So don't bring up CPU usage again. It just makes you look foolish.
<<<<<
lol ok, let's ignore the numbers that anandtech users got, they must just be incapable of properly configuring their systems. Storagereview may just be better at it, so let's assume that all users will be able to configure their systems similarly, even though this is clearly not the case, as indicated by the HDtach numbers.
>>>>>
<< Then you'd have to arbitrarily determine how much disk swapping should be going on, and how much that should be weighted against the rest of the benchmark....you're making it synthetic and not real world. >>
Not arbitrarily at all. We simply write the benchmark to load a set amount of programs and perform a set amount of tasks in those programs. CSA's OfficeBench is already capable of this. Then we record what activities in typical modern computing result in disk swapping and stress those -- like Quake's DEMO1.DEM.
<<<<<
Which programs? Which tasks? Some of these are more disk intensive than others I'm sure, and in different areas as well. They should all be weighted by how liely a given user is to perform those tasks...which is of course impossible to determine for each user because everyone does tifferent things with their computers,and so these decisions become arbitrary. Want to use Quake's DEMO1.DEM? That's ok, but it won't tell you anything about how I benefit from SCSI since I don't play quake.
>>>>>
Regardless, all of my ideas on measuring performance are better than yours. But have it your way. Just eyeball every piece of hardware you evaluate, use the warm fuzzy test to measure the effect a bunch of metal and plastic has on your heart, and then buy whatever makes you happiest. A playful burst of insanity that would bring joy to any asylum.
<<<<<
I have never said that the purchase should be based on emotional feelings, though you love to continually point to that. I have already sstated several times now that purchasing decisions should be based on the best performing device in your price range.
>>>>>
<< The fact that it shows two extremely different drives performing equally shows how ludicrous it is to consider it a valid benchmark. >>
The whole point of my holistic benchmark is to smooth out these aberations in otherwise useful tests.
<<<<<
Our contention is that the test is no longer useful.
>>>>
<< Unlike you Modus, I'm not trying to impose a law or a general rule. >>
Actually, you are. Your rule is: Anything is a good value as long as you think it's a good value; anything is sensible and practical so long as you believe so. Pardon me while I vomit.
<<<<<
Again I say, I am not trying to *impose* a rule. My value rule applies to me, I don't go around telling people they made a foolish decision because they didn't use my model of value. You, OTOH, seem to enjoy telling people that they made foolish purchasing decisions, whnever they choose not to follow your price/performance model.
>>>>>
<< Most upgrades for a home computer could be better spent on other things, such as food >>
Actually, the average Western family consumes far too much food.
<< or simply invested in the bank. >>
Agreed.
<<<<<
So instead of upgrading our computers, as long as they work, all the money that would be spent on computers should go in the bank then right?
>>>>
Not in today's society. Computer literacy is becoming the defining line between the "have" and "have not" careers. A modern computer is an essential household tool for a large bulk of the workforce.
<<<<<
Useful, yes. Essential, unlikely. Basic computing skills can be learned in a classroom and practiced on lab computers. Once you have these basic skills, you may use them in your job which requires them, and do not need a home computer to keep them up. If you want a tech job, then of course a computer would be essential, but a techie is not your average person.
>>>>>
<< I have a friend interning at a company similar to pixar, these so called professionals make very very good money, rendering jobs distributed over their network of powerful workstations take hours. . . obviously performance and reliability means a lot. Do they use SCSI? Hell yeah. Is it foolish for them to do so? Highly unlikely. >>
The bottleneck in unattended 3D rendering is the CPU or video card, not the hard drive. Reliability can be ensured by a cheap IDE RAID server with automatic redundant backups and frequent off site storage. SCSI is not necessary here, nor would it provide a significant productivity boost. The company is obviously able to afford the technology and thinks little of the extra two or three hundred dollars per computer. That doesn't make them foolish, but it's certainly not the optimal use of their funds.
<<<<<
CPU and RAM are a large bottleneck, but once you've got the best there, what's left to upgrade but the disk subsystem? And the disk subsystem is still largely used, just not as much as the CPU or RAM (of course). IDE RAID has extremely high CPU utilization because almost all IDE RAID solutions are software based, they are not hot swappable, so there is more potential downtime.
Considering that every minute wasted costs them quite a bit of money, the extra money was most certainly an optimal use of their funds.
>>>>>
<< Heh, we've been messing around with an embedded system here, the motherboard doesn't have an IDE controller...so I guess IDE ain't free anymore. >>
Embedded systems? Yeah, that really applies to this discussion. Let's bring up Palm Pilots and BlueTooth transmitters while we're at it!
<<<<<
This is an embedded computing system, on an x86 processor (AMD), with SDRAM. It's a small computer, basically. It's using a motherboard, which happens not to have IDE, which means that it's possible to get a motherboard without IDE controllers on it, so youd IDE controller isn't free. I contest your use of the word "Free" anyway, if you purchase a product and it comes bundled with something else, that extra "something else" was not free, it was most certainly included in the cost of your product. For example, snap-on tools offers "free" t-shirts, hats and other such things when you spend a certain amount of money on their tools. They cost twice as much as craftsman tools. You're paying for that "free" stuff, and the brand name.
>>>>>
<< If you are all SCSI with one IDE drive, you can enable one IDE channel, bringing SCSI/IDE systems to 2 IRQs. >>
Which is the same amount in a conventional IDE system, disproving your notion of saving an IRQ. And if you claim you have no IDE devices, then your price/performance ratio is even more absurd because you now have to pay SCSI's cost/megabyte ratio for mass storage. Unless of course you claim you don't require much storage, but that's quite a bit of backtracking, even for you.
<<<<<
Except you now have optimal performance from every device, because you are not limited by the fact that two IDE devices on the same channel cannot communicate simutaneously. If you wish to get optimal performance from an all IDE system, you want every device on a seperate channel, requiring 3 or 4 IRQs, depending on your addon controller.
And no, I do not have mass storage requirements, everything fits quite comfortably on my 18GB drive, and I see no reason why anyone would need a 40GB drive, but if you need that storage then so be it
>>>>>
<< I guess you've never heard of removable hard drive enclosures. Too bad. They're fast and easy >>
Unless you need to read the data on another computer, say, becuase your current one was damaged (which is, oddly enough, a major reason for backups). By far the best backup tool today is a recordable CD drive, whose media can be read by any computer made in the past five years.
<<<<<<
CD-Rs are great for backups except that they are extremely slow. If you have access to another SCSI rig, then you just move your removable drive to that one to get data off of it. Or you move your SCSI card to another computer. If your computer was physically damaged, then you must buy and configure another one anyway...it should be a given that you'd go with SCSI again, and thus as soon as you had a computer to use the data on, you'd have access to such data. In fact, a removable hard drive is great if you've got a computer at home and a computer at work, then you can very easily transport data between them, and if your work computer stops working, you can access everything from your home computer. But now we're getting into an argument over backup mediums, not SCSI vs. IDE.
>>>>This backup argument is now several levels deep. Obviously you've lost every round, <<<<<
lol
>>>>
but it's also important to recall the original problem: your asinine claim that SCSI technology saves time backing up, either due to time saved by not backing up so often (which was explained to you as dangerous regardless of IDE or SCSI) or due to time saved during a SCSI backup procedure (which simply doesn't happen due to the limitations of removable media, and which, furthermore, is irrelevant because backing up is preferably done during system downtime).
<<<<
Backing up from one SCSI HD to another removable SCSI drive is extremely fast, and it bypasses the limitations of removable media. Backing up cannot be performed during system downtime if there is no system downtime.
>>>>>
<< Tell me Modus, how do you write software to access data on the HD without stressing it? >>
Oh brother. First year computer science! You only use the slower system components when necessary. i.e. use a memory array instead of a temporary disk file. This is how all modern software is written: to use the hard drive as little as possible.
<<<<<
lol ok yes, of course the hard drive is not used for a workspace, this is hardly a feature of "modern" software, although I suppose it really depends on what you mean by modern. I assumed you meant within the last few years, not as in dating back to Win3.1. So yes, a program uses RAM for holding temporary data, until there is no more RAM in which case the swapfile kicks in and "stresses the hard drive". Still, starting a program requires meany non sequential reads, and then any program which uses a database is going to frequently hit the hard drive so that no information is lost, and because the database could not possibly fit into system memory.
>>>>
<< Until then I'll shoot down every point he makes. >>
Considering your record thus far, I'd say we'll be here quite a while
<<<<<
Has anyone ever told you that you are pompus and arrogant? Oh, they have? Many times you say? I thought so
I think my record thus far is just fine, thanks.
Facebook
Twitter