Ok, I am back. Or rather, my gf is gone, so I have time for this thread again. Modus, I notice that you have not addressed Radboy or Pariah's last posts, why is this?
>>>>>
<< Modus, if you have tested this software yourself and have seen that it does in fact work, then I will admit that I'm wrong. >>
It's not a question of admiting anything. I'm just trying to point out that SCSI holds no added value as a backup interface for consumer or business PC.
<<<<<
Ok, fine, then I don't admit that I'm wrong, if that makes you happy.
>>>>>
I don't have a rack, but I sometimes sell them with a system so I'll try to remember to check out the various IDE Hot Swap programs.
<<<<<
Please do.
>>>>>
Even so, the fact that so many manufacturers now offer the feature should tell you that it works as advertised.
<<<<<
It's the first I had ever heard of it. Can't be too popular.
>>>>>
If you're going to test it yourself, don't get a Lian Li rack as they seem to be charging very high rates for what is essentially an ordinary rack bundled with Hot Swap software. The ViPowER units look like much bette deals.
<<<<<
I already have a removable rack, but I can't test the software with it here, it's at home. Doubt anyone wants to wait that long.
Even if the IDE racks work with this software, and assuming this software does not introduce any instabilities into the system, there is still the fact that it requires third party software in order to work, which would of course consume additional resources, and do you know if there are any that will run on a Unix machine?
>>>>>
<< Furthermore, please stop calling me Freddie. You are only doing so to be intentionally annoying, and this is highly immature behavior. >>
I'll call you whatever I please, whenever I please.
<<<<<
This is immature and childish.
>>>>>
Besides, it's a term of endearment. And speaking of immaturity, let's not forget your constant, smirking smily faces, your annoying hatchet-job quoting system that waste needless lines, your exageration of my position, etc. There's plenty to go around.
<<<<<
I find it highly unlikely that you are using it as a term of endearment. I fail to see how my method of quoting is immature. Yes, it takes a couple extra lines, and yes, it is usually more inclusive of past discussions, but the purpose is added clarity and a concious effort to keep context clear at all times. How is this immature, or relevant for that matter? If I am guilty of exaggerating your position, then you are at least equally guilty of exaggerating our position, and I don't see how this behavior has been immature.
>>>>>
<< On further investigation about your supposedly hot swappable IDE HD racks, your definition of hot swappable is not the SCSI version of hot swappable. . . I actually have one very similar to this so I know what I'm talking about, I've tried it [in Win95]. >>
Not so fast, Freddie. The rack you tested was not the same as the ones they've had available for the past year or two. I did a little digging of my own, and found out some interesting facts about the latest in IDE hot swap technology:
<<<<<
Actually, the rack I tested was purchased last year, and is exactly identical to the ones you linked to.
>>>>>
- When the manufacturers claim a rack supports 5400rpm or 7200rpm drives, they are lying. They sell the 7200rpm racks for more money because they supposedly have better cooling, but even most 5400rpm units include a small fan which is better cooling than most hard drives receive.
<<<<<
This information is irrelevant to the discussion, and should be taken as a given. Of course the interface is exactly the same for 5400 or 7200 RPM drives.
Anyway, although you have not tested this in order to back up your claims, and although there are still some problems with this solution (like for servers running Unix), I will back down on the hot swappability issue.
>>>>>
Here are just a few of the many examples of such "insulting and demeaning" statements: (Since you have not directly responded to each of these quotes I posted, and since you constantly accuse me of ignoring the smallest minutia of your own recycled posts, I see no problem reposting them once again to see if you will address them.)
<<<<<
I have responded multiple times to each of my quotes. I am not accountable for the others, and I will not attemtp to explain what Red Dawn (not a large participating member), Radboy, Toolman (also not a large participating member), or Pariah were thinking when they said what they did. The fact is that I have backed up those things which you are quoting me as having said, and you have failed to respond to my explanations, instead you simply resort to reposting your "copy and paste" of all our quotes. Hmm, interesting that you criticize me for using copy and paste later in this post, don't you think?
>>>>>
"Skace, you're being largely ignored because you're not exactly contributing much to this thread." -- Fredrick
<<<<<
This I have not yet responded to. I am sorry if this was seen as condescending, I was honestly just trying to be helpful and explain to Skace why he was being ignored.
>>>>>
<< the moment you start to prove him wrong about anything he drops it (i.e. Pariah's question which he childishly still refuses to answer). >>
About exactly what have you "proved me wrong"? It may help your ego to think that or to say so here, and for that I pity you, but in fact you have not made the slightest scratch in my argument.
<<<<<
If you mean that you have not backed down on any points in your argument, this is true. However, I have brought up many points which you have since decided to merely drop instead of continuing to argue. The way I look at it, if you drop an issue, it's as good as admitting you're wrong.
>>>>>
And do you know why? It's not because you're constantly forced to resort to ridiculous imaginary scenarios with no relevance;
<<<<<
As has already been explained to (and ignored by) you, these "ridiculous" scenarios are used to illustrate clearly that there are situations in which your price/performance ratio does not make any sense. Of course the examples are ridiculous, the point is to make it extremely clear.
>>>>>
it's not because you desperately cling to various worthless trifles rather than approach the main points where you've already failed;
<<<<<
Modus, I respond to nearly every line of your posts (something you are certainly not guilty of doing with mine), what points could I possibly be failing to address?
>>>>>
it's not because you resort to cutting and pasting previous posts too absurd to even address and then proudly claiming they've been wrongly ignored.
<<<<<
I "resorted" to cutting and pasting because you asked me to. Somehow you keep forgetting this.
>>>>>
It's because you're wrong, Freddie. My argument makes perfect sense. There is nothing extreme or offensive about it: Modern IDE setups make better practical purchases than modern SCSI setups for business or consumer PC's. Why is that so hard for you to swallow? Even Radboy agrees. Any sensible computer technology observer realizes that, as a value proposition, SCSI holds little weight; its main strongholds are the enterprise server and hardware hobbyist.
<<<<<
Your position is slowly changing and remolding itself. You now admit that SCSI is in fact a good decision in some cases. And you have also toned down your claims, no longer accusing SCSI purchasers of being emotional or irrational. Yes Modus, we are indeed making progress with you.
>>>>>
As for Pariah's "question", I answered it about 100 posts ago. I could repost it for you for the third time, but since you'll probably just ignore the response and continue harping about it, I don't see the point.
<<<<<
His question was a mere yes or no question, about whether you felt the scores were accurate for that drive. You have again ignored his latest post as well. I shouldn't expect anything else from you.
>>>>>
Suffice it to say, WinBench 99 has been proven to be much more than the "completely useless," "totally irrelevant" benchmark you wrongly claimed it to be,
<<<<<
If you want to use Winbench, then you should notice that it shows IDE with twice the CPU utilization of SCSI. When I claimed that IDE had higher CPU utilization, you scoffed at and ridiculed me, so isn't a benchmark that shows those kind of scores irrelevant? Oh wait, would you only like to take the parts of the benchmark that support your opinion? Sure, we can do that for you Modus.
>>>>>
and your effort to suppress a vital truth -- that a top of the line IDE drive could nearly match a SCSI drive on an industry standard benchmark --
<<<<<
Yes, an IDE drive can nearly match a SCSI drive on an outdated benchmark with little or no relevance to the actual performance of a modern disk subsystem. Yes, this benchmark was once the industry standard, and considered by manufacturers to be accurate and valid. Any manufacturer who said "oops, it's not valid anymore" would look like they were backtracking, when the truth is that it is just no longer relevant - they, and the industry, simply prefer to let it slowly fade off into the sunset while bringing in more relevant benchmarks such as IOmeter. So what?
>>>>>
has been revealed as a transparent attempt to censor valid information that would help potential buyers make an informed decision.
<<<<<
I haven't tried to censor anything, nor have I advised any potential buyers to ignore winbench. I have, however, pointed out that the information from winbench is largely invalid.
>>>>>
All participants in this thread have (grudgingly for the SCSI fanatics) acknowledged that the third definition of value, price/performance, is indeed valid to our discussion.
<<<<<
But it is not the only valid one, and it has been demonstrated that it is not always the best one to use.
I can buy a used schoolbus for $1800. It seats 72 passengers. Compare that to a used car that seats 6 people, and the bus is 12x better, right?
It's not relevant that you don't need that kind of "storage" area right? In fact, I say if we're going to compare a bus to a car, and the bus carries 72 passengers, you should find a car that carries 72 passengers as well, so we can have a fair comparison. And don't talk to me about gas mileage, thinking about the future will get you nowhere, I mean gas prices are fluctuating all the time, how could you possibly know if mileage will make enough of a difference to factor in? Acceleration shouldn't mean much either, the bus will get you where you're going, and there can't be that much time lost, I'm sure it only adds up to seconds a day. (you can get the bus here:
http://www.yellowbussales.com/yellowbus/yellowindex.htm - excuse me for not linking)
yes, this is another rediculous example, and no, I do not agree with the above paragraph. It was meant to serve as an illustration of a place where price/performance does not hold up.
>>>>>
This is not dogmatic or difficult to accept. It's the definition of value used by AnandTech's Buyers' Guide,
<<<<<
Note that the buyers guide never says that everyone should buy the "value" system, nor that anyone purchasing anything with higher performance is doing so only for emotional reasons, or that they are foolish for doing so. In fact, the Buyers Guide gladly offers alternate suggestions for those interested in performance systems.
>>>>>
Of the major points you've attempted to make against common sense, Skace or I have debunked every one. In your insistence that we respond to *every* point you think you've made, no matter how incoherent or irrational it actually is, you show how childish you've become.
<<<<<
If you do not understand a point I have made, please ask me to clarify rather than simply ignoring it. If it is irrational, it should be easy to debunk, no need to ignore it. It is not childish to ask that your points be responded to. It is very hard to have a discussion or argument if every point you make is ignored.
>>>>>
If some one accuses Alan Greenspan of using the US Federal Reserve as his personal device to further the control and eventual colonization of Earth by lizard people from the planet Neptune (enslaved under the cruel dominion of Richard Simmons), does he bother to deny it? Of course not. If he were to sincerely deny such a demented argument, people would think he was nuts. But if some one accused Greenspan of caving to Bush Jr's pressure to initiate the deepest tax cuts in history at a point where the government had the opportunity to massively reduce its debt load, he might write a press release in his own defense.
<<<<<
Speaking of ridiculus examples, Modus...
>>>>>
On the other hand, when you concoct absurd fairy tales like a free P133 machine being suitable for modern productivity applications and destroying price/performance considerations because of it, or that StorageReview is incorrect in its well-documented tests that show IDE's processor load equal to SCSI's, I would just as soon not give them the time of day.
<<<<<
Some of StorageReview's tests show equal CPU load (IOmeter), while others do not (your personal favorite, Winbench).
>>>>>
The problem is, you're like a sweating, deranged, Fox Mulder-wannabe reporter in the White House press pit doggedly asking the Press Secretary why Mr. Greenspan has ignored the serious allegations of his collusion with alien life forms controlled by Richard Simmons. You demand a response, regardless of how bizarre your little theories are. And like a tired Mr. Greenspan forced to respond to the manic reporter, I will now address your precious points:
<<<<<
Thank you, oh great master, for stooping down to my level and answering my absurd and ludicrous questions about aliens and mind control. Get real Modus, these are valid points, and the only thing delusional about our discussion are these inane stories you are concocting.
>>>>>
You've shot down nothing but my previous favorable impression of you. And claiming you haven't the time to respond to my arguments doesn't make them false -- it simply makes you too lazy or uncertain to try.
<<<<<
I did respond to your argument, I claimed that I didn't have the time to do it again when I had already done so, hence I copied and pasted, something it has already been shown that you are quite guilty of yourself.
And yes, I'm sorry, but the time that I spend responding to your wonderful posts is indeed limited, I have other interests and responsibilities. I keep this up largely for my own amusement, and because somebody has to keep you in line.
>>>>>
<< RE: "The fact is, as Skace pointed out, the vast majority of what you dredged up had already been put down later in the thread." I reread everything after my posts, and this is simply false. >>
No, it isn't.
<<<<<
Yes it is.
(mom interrupts: Kids, stop arguing or I'll turn this car around!)
Seriously though Modus, put a little more effort into your argument. Back it up.
>>>>>
<< Ah, so now you are qualifying your statements about value with affordability too, huh? >>
I always have. If you can't afford it, you can't afford it.
<<<<<
So the Price/Performance ratio isn't the end all be all rule needing no explanation. You have so far added the following exceptions:
affordability
enterprise level
enthusaist purchases
>>>>>
<< furthermore, you are only using access time, which you have previously ridiculed as being
meaningless on occasion, and didn't you say that it should not be the sole metric for measuring hard drive performance? >>
I have never, ever ridiculed access time as "meaningless". That makes you a liar, Sir Fredrick.
<<<<<
"meaningless on occasion". You said that there were times when access time did not come into play, where STR is the only important factor. Yes, I'm paraphrasing, these are not exact quotes, but that does not make me a liar.
>>>>>
I did, however, show that it should not be used as the sole metric for hard drive performance.
<<<<<
Which is exactly what you went on to do. Hypocritical? No, not at all...
>>>>>
But since access times are where SCSI drives show their biggest advantages, and since the SCSI advocates continually harped on them, and since we've been using them for price/performance comparisons for about three weeks now, I followed with tradition. Rest assured that other benchmarks such as Winbench or STR or IOMeter will yield similar results.
<<<<<
As I have pointed out many times, seek times are not the only place where SCSI drives accel, it's just one of their better areas. If a SCSI drive has high performance (over IDE) elsewhere as well, it should be taken into account, too. You love to use seek time and nothing else, or to use the other differences to say that SCSI performance is between x% and y% better than IDE, where x% is the lowest performance difference on a single metric, and Y% is the highest performance difference. What is far more likely to be the case however, is that both x, y, and everything in between have a cumulative effect (not necessarily x + y mind you) to bring the overall performance difference between an IDE drive and a SCSI drive well above y%.
>>>>>
<< I was unable to find the drives you listed at those prices, but I did find this: cheapest IDE drive -- Generic 15.2GB EIDE 5400RPM, 9MS seek -- $85 -- Cheapest 7200RPM IDE drive -- Maxtor 15.2GB EIDE 7200RPM, 9MS seek -- $96 -- Better scoop up that generic drive. >>
First you imply that I made up the original prices, which anyone can easily find here and here,
<<<<<
I was not implying anything. I clearly stated that I had been unable to find those drives for those prices. This was a true statement. You have since provided links to them, and this is appreciated. For future reference, is would be better if you backed up or clearly stated where you got numbers from, rather than forcing us to ask for verification.
>>>>>
and then you use try to present your own ridiculous comparison as in any way useful? That's low, even for you, Freddie.
<<<<<
Speaking of implications, you imply that I typically am willing to stoop to low actions/arguments with this statement. My comparison was done very quickly, and may not have been ideal in any sense, but it illustrated that it is possible to show a skewed price/performance ratio very easily.
>>>>>
Regardless, your comparison is stupid because the listed seek times are manufacturers' ratings and thus estimated or exaggerated. Further, one drive is listed as "generic", so its true performance can never be confirmed by a reputable site. Thus price/performance cannot be judged.
<<<<<
If you are truly interested, you could contact the store, they may have information on that.
>>>>>
<< Ever heard of rtf or csv? Both versions of office support this seamlessly. >>
"Seamlessly," LOL . Have you ever tried to get Word features to survive a conversion to RTF format intact? Mail merges? Headers and footers? Floating pictures? Footnotes and Endnotes? Embedded hyperlinks? Date, time, and filename stamps? Hanging indents? RTF is as incapable of addressing those word processing needs as a free P133 is of Office 2000 productivity.
<<<<<
Admittedly, I do not use most of those features, but neither do most businesses. And certainly on a home conputer, it's unlikely that you will be using most of these features. A business I have occasion to work for does use ACT! and Word 97 for mail merges, and while this is unrelated to rtf, one of their computers is an AMD 486 133MHz. It does not choke on these tasks.
<< It's perfectly capable of editing a spreadsheet, or creating a document in word. >>
>>>>>
Sure, but what kind of spreadsheet, and what kind of document? Anything beyond "the quick brown fox jumped over the lazy dog" or Sunday's To Do list is going to seriously tax the machine's resources.
<<<<<
Speaking of that famous sentence, open (or create) a word document, and type in:
=rand()
on a line by itself.
Irrelevant? yes. I found it to be interesting anyway, and if you don't appreciate this little side track, I apologize in advance.
Back to the topic at hand: unless you are doing some serious number crunching, an older machine will handle it fine. Even with major complex calculations, the machine will not simply stop working, it will just take longer to perform the operations.
As far as word processing goes, the AMD machine has been used in part for compilation and editing of a 600+ page manual, with no problems. Not the fastest machine in the world, but usable by the employees for the necessary tasks.
>>>>>
Throw in a few concurrent programs like ICQ or Winamp, an open browser window, and couple that with the low memory typically found in those systems (anything more than 64M on most P133 compatible chipsets resulted in decreased performance due to L2 cache mapping techniques) and you've got a recipe for disaster. Illegal operations, sluggish window navigation -- basically, an impossible situation for anyone trying to get real work done.
<<<<<
If you're chatting on ICQ, you're not trying very hard to get real work done, and if you're in a professional situation, chances are you won't be running Winamp either. Also note that a slow hard drive will have an even greater negative impact on such low memory systems. Nevertheless, illegal operations are rarely the fault of a slow processor, but rather poor programming. Sluggish window navigation does not necessitate a faster system. Do you "need" to have instant window navigation? Or is it your emotional desire to have a system which meets your expectations for performance even if they are less than the bare minimum?
>>>>>>
Look, much as you'd like to trumpet the stolen Pentium 133 as the killer bullet of an entire economic theory, it's not going to happen. First, such machines are not free any more than a Duron 800 is free, and second, they are inadequate for today's productivity needs.
<<<<<<
I have recieved several older systems for free, and they were not stolen. However, if you don't know anyone looking to get rid of an old computer, you can purchase one at a flea market for as little as $20 (without a monitor).
>>>>>
<< If installed correctly, Office 2k can actually be less bloated then 95 or 97, believe it or not. >>
Um, not. Hard drive space is not the critical bloat factor. Program resources and code bloat are the things that actually affect performance. In that regard Office 2000 is more bloated than any previous Microsoft suite -- regardless of whether you install the clipart or not.
<<<<<
I found that Word 2k used less resources on my machine than word 97 did, it was also quicker and more responsive. Of course, I had turned off all the features I did not need, such as the annoying paperclip.
>>>>>
<< RE: Office 2000, video editing, speech recognition, CAD, 3D animation. A P133 can perform these tasks. Slowly. But it CAN perform them. >>
Wrong. Speech recognition packages, for instance, require at least a 300 MHz processor.
<<<<<
I had an old Dragon naturally speaking package running on my P133. I never used it much, but it did work. Granted, it was annoying to use and slow, but it did work.
>>>>>
Working with complex 3D animations is laughable on anything before the AGP era of faster video cards.
<<<<<
But can be done if your time is worthless.
>>>>>
I could go on, but you get the idea. By no reasonable standard could a P133 machine be considered suitable for modern productivity applications.
<<<<<
Your average user will not be running the applications which perform video editing, speech recognition, or CAD, so this is irrelevant. Even so, I have already provided arguments above.
>>>>
<< If you believe that standing still is progress, it's no wonder you love IDE. >>
We'll have to add this to the list of "insulting, demeaning" comments you and other SCSI advocates have made.
<<<<
What's funny and sad at the same time, is that you actually keep a list. And before you start trumpeting this as well, I was joking. This should be obvious to all who read my post.
>>>>>
<< For what I do with my computer, it was money well spent. >>
Could the money have been spent even better? Sure: IDE.
<<<<<
Not for what I do with my computer, not for the performance level I expect from my computer, not for the time savings with SCSI, no. IDE would have been a poor choice.
>>>>>
<< Nothing wrong with being generous, but I'm not buying anyone a car while I sit at home suffering with my sluggish all IDE system. >>
No, you'd rather have SCSI and drive a Ferrari. Very practical.
<<<<<
Actually, I'm not a fan of sports cars. But I'm not going to be buying a bus any time soon either.
>>>>>
<< I was not "forced" into anything >>
Sure you were. You forced yourself to buy the best you could afford because you desired maximum performance and decided to pursue it.
<<<<<
It's forcing in so much as you force yourself to purchase less than the best in order to obtain the maximum price/performance ratio. Otherwise, it's known as a decision. I decided to buy the best I could afford within my budget ($1500) for a new computer.
>>>>>
<< The idea was to do it right the first time and save myself the hassle and expense of upgrading later, while being able to enjoy the performance right away. >>
This is an example of a seemingly logical, yet deeply flawed buying strategy. I call it "obsolescence denial". The problem is that it can work so well for some products but so poorly for others. [insert silly washing machine analogy here]
But this kind of thinking belongs nowhere near the computer industry. Computer hardware is now a disposable commodity.
<<<<<
No wonder you advocate purchasing whatever is cheap. Hardware is not always a disposable commodity. I expect my computer to last me a long time without necessitating many major upgrades along the way, because I do not always have the money to spend on computer upgrades.
>>>>>
The technology advances so quickly that the future is too nebulous to even anticipate. Only the present is at issue because only the present can possibly be understood or controlled. To buy any computer hardware today and give the slightest thought to its longevity is to deny the planned obsolescence that keeps this industry vibrant. Whether you upgrade in one year or two is irrelevant. All that matters is a product's price/performance ratio today. Let tomorrow bury today.
<<<<<
And yet this method has worked so well for me; my system is still going strong after 2 years, and I have not grown tired of it (or become frustrated with it's speed) - it can still perform every task I throw at it, with no problem or hesitation. Whether I upgrade in one year or two is relevant to me, the longevity of my system matters. I may have spent somewhat more up front (although with shopping around and price comparisons, the price difference turned out to be rather miniscule), in the long run I have saved by keeping the same system for a longer period of time, and in the mean time I have been more satisfied with my computer, because it was built to handle any task I might throw at it, and it has done just that.
>>>>>
<< The fact that anandtech users are apparently incapable of configuring their IDE systems properly shows that in theory, CPU utilization is very close, while in practice, it is difficult to configure properly. >>
Not at all. It simply shows that StorageReview's testing procedures are consistent and methodical, whereas the AnandTech users who participated in an impromptu survey had not previously taken the same care to ensure their systems were running at peak efficiency. The SCSI users, being naturally more committed in this area, did.
<<<<<
There is relatively little to be done with a SCSI system to ensure optimal configuration - if it's working, it's probably optimally configured. IDE users who were seeing high CPU utilization were quite distressed about this, and did seek help in configuring their systems optimally. Surely you don't think that SCSI users are more technologically advanced and more capable at configuring their systems correctly.
>>>>>
You're really pulling teeth here. The ONLY things required for similar results to the 0.6% CPU usage tested by StorageReview are the latest IDE busmaster drivers (always necessary for maximum performance and compatibility, and these days included in the VIA service pack at no extra clicks) and a clean System Tray. This is not difficult to achieve. Again, every single qualified expert agrees that IDE busmaster systems now match SCSI in CPU usage.
Why don't you?
<<<<<
Actually I have read a lot of discussion about this issue, and it has not been resolved completely. Some benchmarks show that CPU utilization is the same, while others show that it is different, including benchmarks used by StorageReview. If StorageReview explicitly states that there is no difference, then that is a very valued and well respected opinion on the subject, but not the only one.
>>>>>
<< If you are going to throw numbers around, it's nice if you can back them up. Where do you get 0.60% from? >>
You know exactly where I got the number, you're just using your old deception of implying that I make numbers up. You should know by now how meticulous I am when quoting figures and researching text.
<<<<<
Funny, you'd think you'd be meticulous about noting where you got them from. I actually did not know exactly where you got that number from, which is why I asked. I am not using any underhanded tactics to try to discredit you, I am merely requesting that you back up figures when you are throwing them around.
>>>>>
The 0.6% processor usage is actually on the high side of the Threadmark and Winbench tests from StorageReview on IDE drives. Look them up yourself. What's even funnier is that, upon switching to the Windows 2000 testbed (with its ACPI IRQ sharing that you claim reduces system performance) the CPU usage of IDE drives actually went down.
<<<<<
You don't know whta their Win2k setup was like, and you don't know if the IDE IRQs were shared. Nor were they taxing the whole system, so if the lack of IRQs forced two other devices to share, thus degrading their performance, it would not be apparent.
>>>>>
<< Further, storagereview specifically said that each additional channel uses more resources, not each additional driver. >>
This is incoherent babble. One driver, two channels.
<<<<<
They said it, not me. I'm sorry you couldn't understand it because it was perfectly clear to me. Storage review said that each additional channel uses more resources. If you think they meant driver, take it up with them, ask them to reword it.
>>>>>
<< If you apply this purchasing strategy to everything, you've got a 24% slowdown here, a 2.4% slowdown there, a 1.87% slow down there, a 5% slowdown here...it adds up to something that's a whole lot slower. >>
It does not. You're just wrong on this, Freddie, and we've told you before. You cannot add the individual factors that may or may not affect performance into a big lump sum and declare the result the actual performance difference. That leads to situations where a part is 10% slower in ten different areas so is judged 100% slower -- logically and mathematically impossible. The performance difference between IDE and SCSI has already been measured.
<<<<<
I agree that they cannot be simply added together, but I am contesting your methods: either average out the difference, or throw out all but the highest and lowest differences. That certainly does not give you an accurate picture of the slowdown experienced. The slowdowns actually do accumulate, but not necessarily in a manner such as 24% slower + 2.4% slower + 1.87% slower + 5% slower = 33.27% slower -- I am not proposing this and nowhere have I said that this is the case.
>>>>>
<< I already showed how numbers could easily be manipulated to show that SCSI is a better value. >>
Manipulated, sure. Anything can be manipulated!
<<<<<
My point exactly. So your particular manipulation of the numbers to prove your point is meaningless.
>>>>>
That's like saying the P4 is cheaper than the Athlon because we can "easily manipulate numbers" to show it. What a retarded statement. Anyone representing the numbers truthfully sees that IDE is a better practical value. And you know it.
<<<<<
It depends on what you're doing.
>>>>>
But then, Freddie's definition can mean anything Freddie wants whenever Freddie wants it. So who could argue with that?
<<<<<
I've already defended this elsewhere, but in true Modus style it's been ignored because it's easier for you to say that I believe that value is nothing more than whatever I feel like purchasing at the moment.
>>>>>
<< Apparently they are contradicting themselves. >>
Nope. Their only statements against Winbench have been to point out its foibles with the Seagate X15 and a couple specialized hardware RAID controllers. It doesn't change their stance on the benchmark as a whole.
<<<<<
"there are some legitimate concerns raised that WinBench, while accurate for testing the application workload it purports to measure, provides too light of a load on the system to represent performance on a more general basis. Further, as a given release of WinBench ages, manufacturers become better at "tuning" drive performance to reflect high numbers without a corresponding increase in actual performance."
>>>>>
If it did, why would they keep using it?
<<<<<
Standardization. This has been pointed out long ago.
>>>>>
Why would they not strike the records from their archives?
<<<<<
This would require that they retest every old hard drive in order to remain consistent, which would be a pointless waste of time on their part.
>>>>>
Why would they not change their official disclosure statement which they list with every article and which they update quite often, to reflect their supposed new wisdom?
<<<<<
I don't know, but I found this in the review of the Seagate 18XL:
Many readers are likely interested in comparisons between the Cheetah 18XL and StorageReview.com's current leaderboard champ, the Quantum Atlas 10k II. Despite its gains over the 36LP, the 18XL simply can't keep up with the Quantum drive when it comes to WinBench 99. In such a comparison, the 18XL falls behind the Atlas by margins of 15% and 25% respectively in the Business and High-End Disk WinMarks.
However, everyone knows that these days, IOMeter is where the action is here at StorageReview.com. Let's see how the 18XL stacks up in this more important test.
While we're on the topic of StorageReview, I also found this in the reference section:
Performance: The hard disk plays a very important role in overall system performance, probably more than most people recognize (though that is changing now as hard drives get more of the attention they deserve). The speed at which the PC boots up and programs load is directly related to hard disk speed. The hard disk's performance is also critical when multitasking is being used or when processing large amounts of data such as graphics work, editing sound and video, or working with databases.
Reliability: One way to assess the importance of an item of hardware is to consider how much grief is caused if it fails. By this standard, the hard disk is the most important component by a long shot. As I often say, hardware can be replaced, but data cannot. A good quality hard disk, combined with smart maintenance and backup habits, can help ensure that the nightmare of data loss doesn't become part of your life.
I actually haven't come across their disclosure statement, but I admittedly haven't looked very hard.
>>>>>
<< if I disagree with the rule that "Modus's definition of value applies to everything and everyone, always". . . >>
Since I have never proposed such a rule, your statement is moot. The appropriate definition of value applies only to those purchases intended to be sensible and practical, an optimal use of funds. Other goals, such as maximum performance within a specific budget or maximum enjoyment, are obviously incompatible with that definition of value.
<<<<<
And since they don't meet your definition, you have repeatedly insisted that they must be foolish or emotional purchases. Our whole point is that it is perfectly reasonable to make a purchase based on maximum performance within a specific budget.
>>>>>
<< Guess what, a faster hard drive will allow all of these applications to start up more quickly >>
That does not make disk performance the limiting factor, as you mistakenly claimed. The limiting factor in any performance measurement is the factor that most affects performance. Disk performance is not the most important factor in any of today's common productivity applications.
<<<<<
See my storage review quote above. It may not be the limiting factor in netscape's performance, for example, assuming that you have plenty of memory so that your computer isn't hitting the swap file while you browse, and assuming that you have no disk cache. Otherwise even then it will play a role in the performance, though it will probably not be the main limiting factor. For many database applications (which would be used by businesses, for example) are constantly reading from and writing to the hard drive, in which case it can often become the limiting factor. If you are multitasking, it becomes even more of a limiting factor.
>>>>>
For example, if a Photoshop filter takes 3.0 seconds to process on an IDE drive which needs a 0.009 second seek to retrieve the data, while a SCSI drive needs only 0.005 seconds, the Photoshop filter still takes 2.994 seconds. The time saved is a blink of an eye. And now imagine we had a magical SCSI drive with 0.000 second access times. How long would the Photoshop filter take? Exactly 2.991 seconds. Other productivity tasks follow a similar model -- SCSI speeds up only the portion of the task that waits for the disk, nothing else.
<<<<<
Assuming that these operations never cause the system to use the swap file. Working with multiple large images with several levels of undo can easily cause swapping, especially in Windows 2000, which will use the swapfile even if there is room in main memory.
>>>>
<< Considering that STRs are no longer an issue with most modern hard drives, we can focus on the more important issue, which is latency. >>
We can't just ignore STR! It's a part of real world performance. You can't throw it away just because IDE drives match SCSI drives there. You have to include it to some degree in any holistic performance comparison.
<<<<<
This has already been addressed.
>>>>>
They are not. See above. But don't believe me, ask around this forum. Consult experts. Send some emails. Performance factors are NOT cumulative. Overall performance measurements stands on its on. A video card with a 10% better fillrate and 10% higher clock speed and 10% faster T&L unit is not 30% faster. That's lunacy.
<<<<<
Nor is the video card merely 10% faster. This is the point I have been trying to make.
>>>>>
To find out the real performance increase, you either run a benchmark that stresses the whole card, or you average the different primitive factors into a weighted figure. With StorageReview's drive database, we can do both. The important benchmarks they run are IOMeter, Winbench 99, STR, and access time. We simply average those four according to whatever scheme you think is fair (ie. not 100% access time or 0% Winbench) and come out with a very rounded measure of holistic drive performance -- one that does not completely ignore STR.
<<<<<
Modus, say we're using a benchmark which uses a system of arbitrary points. Now let's say that the IDE and SCSI drives have identical STRs, which causes them to recieve an equal number of points in that area, let's say 200 for IDE and 200 for SCSI. They have effectively cancelled eachother out. Regardless, SCSI drives still have higher STRs, so including STRs works in SCSI's favor.
>>>>>
The burden is on you to provide proof that CD to CD burns are not possible on the same IDE channel. Clearly, they are possible, and it's just more anti-IDE scare tactics on your part. Besides, optimal drive placement will make the issue moot to begin with.
<<<<<
I said that it may be impossible, not that it will be. My experience has been that sometimes it will work, and sometimes it will not. Depending on the system, and what else the user is doing at the time of burning. My proof that it may not be possible is anecdotal, but that is all that is necessary. I have helped friends with 3 IDE systems trying to get their CD burners to burn CD-to-CD, with both on the same channel. At best they were intermittently successful, so my advice to them was to copy to the hard drive first.
As far as optimal drive placement, you must have a poor memory Modus (worse than mine, and that's pretty sad), becasue the point I originally brought up, and the whole reason that we are discussing this, is because I pointed out that optimal drive placement can be quite difficult in certain configurations.
>>>>>
<< SCSI HD -> SCSI HD will be faster than IDE HD -> IDE HD >>
But SCSI 12x is no faster than IDE 12x,
<<<<<
Unless you're relying on burnproof for the IDE drive.
>>>>>
and CDR is the most flexible backup tool. For removable hard drives, STR is the limiting factor in drive image backups, and newer IDE drives are rivaling SCSI in that category. Again, SCSI backup systems hold no significant speed advantage.
<<<<<
In drive imaging, STR plays a large role, but is not the only factor. Seek times and buffers also play a role, as well as overall bandwidth and multitasking ability (remember, SCSI drives can be on the same channel, talking to eachother, with no problem)
>>>>>
<< RE: "the IDE guy plugs it in and the motherboard detects it in two seconds." SCSI guy plugs in SCSI card, motherboard detects it in two seconds. Time difference? None. >>
Pure bull. I challenge you, Sir Fredrick, to a race. I will race for IDE and you for SCSI. You will install the controller and drive, while I will install only the drive. Care to wager on the outcome? <<<<<
Modus, I have done this before. The most time consuming part is the physical installation and plugging in the cables. Transplanting from one system to another is a piece of cake. Where I would have to plug in a card, you would have to plug in a cable. Once that is done, the computer may be started. With a SCSI setup, there is no need to enter the BIOS with the new drive installed. The SCSI card will be recognized by the BIOS automatically, the SCSI BIOS will do it's thing (and will not need to be configured at all since it was already operational in another system), the OS will detect the card and drive and install the drivers for them (if this is an Adaptec card, no intervention here is necessary), and depending on the OS, it will require a restart.
For an IDE drive, you can just plug it in, assuming the secondary channel is not being used, and turn the computer on. Once the computer is on you must enter the BIOS, and configure the drive, or set it to auto detect (which makes booting slower), save to BIOS and restart the computer. Windows will start up and detect the drive, and depending on which version you are using, it may require a restart.
Time of installation between the two is very similar, and would be more influenced by things such as who is faster with a screwdriver. However, given that those times would be identical, I would wager that it is probable that SCSI would in fact be ahead, though this is really a moot point in my opinion, because it is unlikely that there would be more than a minutes difference between the two.
>>>>>
<< Fact is, IRQ sharing reduces performance, and can cause conflicts with some devices. Yes, in Win2k. Have I said this before? Yes. >>
Were you as wrong then as you are now? Yup.
<<<<<
Have you proven me wrong? Nope.
>>>>>
<< you are also acting as an antagonist >>
It comes down to this, Freddie: You admit that IDE drives have a higher price/performance ratio than SCSI drives. You also admit that one dictionary definition of value -- price/performance -- is valid. The only thing left for you to admit, for me to stop "antagonizing" you, is that this definition of value is the most efficient way to make a sensible, practical purchase.
<<<<<
The only thing antagonizing is your insistance on calling me Freddie, which you have been doing for the sole purpose of annoying me, even after I politely requested that you stop.
Ok, that's enough for me tonight.
Facebook
Twitter