I ran Scandisk. Before running it I noticed things are doing better. The pointer did not give me the finger for a while. Now it seems to be working normally. I also adjusted the graphics card so it would do better with performance instead of quality. Still want to see how the other hard drive performs in comparison. Thanks for all the help.
Hard Drive questions?
- Thread starter MONKEYMAN
- Start date
- Status
No worries!
Originally Posted By: StevieC
tropic, while that is a good idea, I have seen bad drives that past the sea-tools and other MFG's tests that are in the stage of going bad or have controllers that are failing but good enough to fool the test. Just something to keep in mind.
Thanks for correcting me. Looks like SeaTools is just a suite of HDD fitness tests now. MaxBlast is Seagate's HDD cloning utility: MaxBlast. Guess I'm getting old.
tropic, while that is a good idea, I have seen bad drives that past the sea-tools and other MFG's tests that are in the stage of going bad or have controllers that are failing but good enough to fool the test. Just something to keep in mind.
Thanks for correcting me. Looks like SeaTools is just a suite of HDD fitness tests now. MaxBlast is Seagate's HDD cloning utility: MaxBlast. Guess I'm getting old.
tropic not getting old, you are getting better with age like a fine wine!
It's all good, I used to work full time with this stuff and so I have seen a lot of weird problems. I still keep current by fixing a lot of friends/family stuff etc.
It's all good, I used to work full time with this stuff and so I have seen a lot of weird problems. I still keep current by fixing a lot of friends/family stuff etc.
My landlord is coming back tomorrow. I will try the MaxBlast when I get some time. Speaking of age, what I get to look forward to in the near future is regular colonoscopies. Guess it is a right of passage of old age. I learned a lot here. Thanks again!
The main performance parameter that can make a hard driver faster than the other is linear bit density (how many bits per track), then linear track density (how many track per inch). Since almost all HD are now 7200rpm (on desktop), the bits per track gives you the biggest improvement in speed. Track per inch gives you less but still significant amount (since the servo takes time to settle if you have high density, so it is not a 1:1 improvement like bit per track does).
For a same type of drive (i.e. the 40GB is not a 10k rpm monster and the 160GB is not a 5200 rpm slow poke, both are 7200 rpm), switching from a 40GB to 160GB usually give you a 4x density improvement if the platter count is the same. For the same generation of drive, 40GB vs. 160GB will not yield a huge improvement (1 platter vs 4 platter).
It is more than likely a 40GB HD has 2MB of buffer while the 160GB has 8MB, the performance difference between 2 to 8 is huge, but from 8 to 16 or 16 to 32 is minor.
Also, if you fill the 40GB to 50%, it is going to be much much slower than having used only 12.5% of a 160GB. File system tends to keep stuff together starting at lower LBA (outer edge of the platter), so the bits per track is higher (therefore faster) and the distance to seek the head is shorter (1/16 of the radius vs 1/3 of the radius, for example).
Finally, for most PC that has a modern processor like P4 and Sempron, the bottleneck after the net connection speed is the ram size, then HD speed.
For a same type of drive (i.e. the 40GB is not a 10k rpm monster and the 160GB is not a 5200 rpm slow poke, both are 7200 rpm), switching from a 40GB to 160GB usually give you a 4x density improvement if the platter count is the same. For the same generation of drive, 40GB vs. 160GB will not yield a huge improvement (1 platter vs 4 platter).
It is more than likely a 40GB HD has 2MB of buffer while the 160GB has 8MB, the performance difference between 2 to 8 is huge, but from 8 to 16 or 16 to 32 is minor.
Also, if you fill the 40GB to 50%, it is going to be much much slower than having used only 12.5% of a 160GB. File system tends to keep stuff together starting at lower LBA (outer edge of the platter), so the bits per track is higher (therefore faster) and the distance to seek the head is shorter (1/16 of the radius vs 1/3 of the radius, for example).
Finally, for most PC that has a modern processor like P4 and Sempron, the bottleneck after the net connection speed is the ram size, then HD speed.
The RAM on the Pentium 4 is 2gb. The ram on the Sempron 3000+ is 750 mb with a newer Seagate 80 gb Sata drive. Both use the same connection. I have AT&T Elite on a 2Wire Gateway. The big difference is it takes longer to load/open AOL on the Pentium 4 and the Sempron starts the AOL program faster and is ready to use quicker after booting up. It will be interesting to see how the 160 gb Seagate drive will work with the Pentium 4. PandaBear, thanks. You seem to have mastered a lot of info.
I had a similar experience yesterday with a paying customer. Something they broke, by messing with things they shouldn't have been messing with netted them two storage arrays (about 5TB of storage) that no longer had any useful data on them.
All because they wanted to save a few hundred dollars and not call the vendor, us!
Well, we've been billing at $275/hour for the past two days to fix their money saving service activity and when we are done, they'll have to recover from tape.
Of course, they want things done yesterday, complaining about the response, citing their contract. I just remind them, as politely as possible that this is a mess of their own creation, not covered by the contract, and that I can concentrate on finding a solution much better and probably faster if my phone doesn't ring every 5 minutes by some new manager calling me.
Originally Posted By: simple_gifts
Quote:
Lastly, how do you guys deal with those people that become overwhelmed and defensive with anything computer related?
Perhaps you need to frame their situation differently... Do the work and use general terms to explain what you are doing. I have the downfall of explaining exactly what the issue; I need to stick to something like: You have issues with the security of your computer. We need to work thru the issues to get you back running correctly. I'm going to need to do several things.... and so on. Using "We" usually helps, connoting a team effort. When someone crashes in and starts mucking with the "P" of a PC, people can get territorial.
If not...
do the minimum, pack up, leave; identify more appreciate pro-bono candidates.
It may be abrupt but doing $50-$75/hr work (for free) that is occasionally aggravating enough without additional emotional stress.
I work in IT (don't do PC work) and have spent 10s of HOURS fixing people's computers and educating them. When my advice is not heeded it smacks of a professional insult, and consequences of not heeding my advice are no crisis in my mind.
Plenty of appreciative people out there
All because they wanted to save a few hundred dollars and not call the vendor, us!
Well, we've been billing at $275/hour for the past two days to fix their money saving service activity and when we are done, they'll have to recover from tape.
Of course, they want things done yesterday, complaining about the response, citing their contract. I just remind them, as politely as possible that this is a mess of their own creation, not covered by the contract, and that I can concentrate on finding a solution much better and probably faster if my phone doesn't ring every 5 minutes by some new manager calling me.
Originally Posted By: simple_gifts
Quote:
Lastly, how do you guys deal with those people that become overwhelmed and defensive with anything computer related?
Perhaps you need to frame their situation differently... Do the work and use general terms to explain what you are doing. I have the downfall of explaining exactly what the issue; I need to stick to something like: You have issues with the security of your computer. We need to work thru the issues to get you back running correctly. I'm going to need to do several things.... and so on. Using "We" usually helps, connoting a team effort. When someone crashes in and starts mucking with the "P" of a PC, people can get territorial.
If not...
do the minimum, pack up, leave; identify more appreciate pro-bono candidates.
It may be abrupt but doing $50-$75/hr work (for free) that is occasionally aggravating enough without additional emotional stress.
I work in IT (don't do PC work) and have spent 10s of HOURS fixing people's computers and educating them. When my advice is not heeded it smacks of a professional insult, and consequences of not heeding my advice are no crisis in my mind.
Plenty of appreciative people out there
The Seagate hard drive is working well. The computer is running fast. I ended up just installing Windows from scratch since I was not sure I would have a good image from the other hard drive since it started hanging up again. Anyway, I have spent close to a month on this project and the somewhat funny thing is she likes the computer I built better than her Dell. So the conversation she had with me was "why do I need the other computer, the one I am using is working just fine". I said it was not always going to work fine, so I would continue to get her other computer ready. Thanks again for all the help!
Originally Posted By: MONKEYMAN
The RAM on the Pentium 4 is 2gb. The ram on the Sempron 3000+ is 750 mb with a newer Seagate 80 gb Sata drive. Both use the same connection. I have AT&T Elite on a 2Wire Gateway. The big difference is it takes longer to load/open AOL on the Pentium 4 and the Sempron starts the AOL program faster and is ready to use quicker after booting up. It will be interesting to see how the 160 gb Seagate drive will work with the Pentium 4. PandaBear, thanks. You seem to have mastered a lot of info.
I used to write firmware for Maxtor before Seagate took over.
The RAM on the Pentium 4 is 2gb. The ram on the Sempron 3000+ is 750 mb with a newer Seagate 80 gb Sata drive. Both use the same connection. I have AT&T Elite on a 2Wire Gateway. The big difference is it takes longer to load/open AOL on the Pentium 4 and the Sempron starts the AOL program faster and is ready to use quicker after booting up. It will be interesting to see how the 160 gb Seagate drive will work with the Pentium 4. PandaBear, thanks. You seem to have mastered a lot of info.
I used to write firmware for Maxtor before Seagate took over.
Originally Posted By: PandaBear
I used to write firmware for Maxtor before Seagate took over.
I've always been meaning to ask someone like you: As firmware is architecture-specific (and assuming that odds of switching architecture is low), do you folks write firmware in Assembly?
I used to write firmware for Maxtor before Seagate took over.
I've always been meaning to ask someone like you: As firmware is architecture-specific (and assuming that odds of switching architecture is low), do you folks write firmware in Assembly?
Maxtor's firmware was mainly in Assembly, except the self test in manufacturing process. It was manageable at first but all the new features, changes make it unreliable all the time, delaying product shipment, lowering yield, etc. Many of Maxtor's quality problem was due to this legacy architectures that was in use for more than 10 years. Yet for some politics and cost reasons, they continue to use it till the day Seagate took over.
Seagate's firmware switched to C in the mid 90s. It is much slower at first but as processors power cost less, the benefits really show in how quickly you can fix a problem and pass knowledge from one engineer to another, reduce overhead, and therefore increase reliability.
Quantum prior to Maxtor's merger had a C architecture as well, but Maxtor threw it out as all Quantum drives were made by Panasonic, therefore more expensive. A lot of the engineers went to a company call Marvell and started an ASIC/firmware architecture that is very similar to Quantum's previous architecture, and is now one of the architecture used by Samsung and Western Digital.
Western Digital has a philosophy of 2 parallel architecture overlapped at the same time. So that in case one has problem they have a second source. This cost more but they seems to think that it is worth it. One of their architecture is Marvell, the other I'm not sure. Same for Samsung, their hard drive used to be junk until they started using outside components like the Marvell.
Hitachi was all assembly until they bought IBM's drive division. IBM used to be C for a while and the engineers from those C architecture helps transformed some of Hitachi's architecture gradually to C. I was told by some of my friends that they are in the process of merging the 2 sides into C.
Seagate's firmware switched to C in the mid 90s. It is much slower at first but as processors power cost less, the benefits really show in how quickly you can fix a problem and pass knowledge from one engineer to another, reduce overhead, and therefore increase reliability.
Quantum prior to Maxtor's merger had a C architecture as well, but Maxtor threw it out as all Quantum drives were made by Panasonic, therefore more expensive. A lot of the engineers went to a company call Marvell and started an ASIC/firmware architecture that is very similar to Quantum's previous architecture, and is now one of the architecture used by Samsung and Western Digital.
Western Digital has a philosophy of 2 parallel architecture overlapped at the same time. So that in case one has problem they have a second source. This cost more but they seems to think that it is worth it. One of their architecture is Marvell, the other I'm not sure. Same for Samsung, their hard drive used to be junk until they started using outside components like the Marvell.
Hitachi was all assembly until they bought IBM's drive division. IBM used to be C for a while and the engineers from those C architecture helps transformed some of Hitachi's architecture gradually to C. I was told by some of my friends that they are in the process of merging the 2 sides into C.
Originally Posted By: uc50ic4more
Originally Posted By: PandaBear
I used to write firmware for Maxtor before Seagate took over.
I've always been meaning to ask someone like you: As firmware is architecture-specific (and assuming that odds of switching architecture is low), do you folks write firmware in Assembly?
Most firmware is in C due to maintainability and low performance hit. Compilers are so good nowadays that it is hardly any worse than hand written assembly. The only parts still in assembly is the very low level drivers dealing with boot up, switching between modes, power savings, and preparing the memory map so that C can function (i.e. loading the initial code into memory, like the BIOS of our PC).
Architecture is redone all the time, mostly because every generation of firmware needs to work differently and we patch the architecture up to a point that it is either very inefficient, too unreadable, or just simply the way data flows have to change in newer products. The trend in the past was small code size (due to limited memory) and recently it is multi-threading (quicker response to input).
The guideline is we always write code for people, then machine. What it means is it must be easy to read and not confusing, so you can fix problems and make changes easily. The size and speed of the code is not as important. In the worst case you can always run the chip a little faster or buy a little more RAM, but if you miss the product's shipment by a week, you probably loses the most profitable week of the product's release.
Originally Posted By: PandaBear
I used to write firmware for Maxtor before Seagate took over.
I've always been meaning to ask someone like you: As firmware is architecture-specific (and assuming that odds of switching architecture is low), do you folks write firmware in Assembly?
Most firmware is in C due to maintainability and low performance hit. Compilers are so good nowadays that it is hardly any worse than hand written assembly. The only parts still in assembly is the very low level drivers dealing with boot up, switching between modes, power savings, and preparing the memory map so that C can function (i.e. loading the initial code into memory, like the BIOS of our PC).
Architecture is redone all the time, mostly because every generation of firmware needs to work differently and we patch the architecture up to a point that it is either very inefficient, too unreadable, or just simply the way data flows have to change in newer products. The trend in the past was small code size (due to limited memory) and recently it is multi-threading (quicker response to input).
The guideline is we always write code for people, then machine. What it means is it must be easy to read and not confusing, so you can fix problems and make changes easily. The size and speed of the code is not as important. In the worst case you can always run the chip a little faster or buy a little more RAM, but if you miss the product's shipment by a week, you probably loses the most profitable week of the product's release.
Originally Posted By: PandaBear
Most firmware is in C due to maintainability and low performance hit. Compilers are so good nowadays that it is hardly any worse than hand written assembly.
Thanks for clearing that up. I had heard here and there that a good compiler can rival the performance of assembly, but I ain't gots the learnin' to knows any better. All of my language knowledge is in interpreted languages, so anything at that low a level is akin to black magic to me. I have looked at Assembly code before, and it looks very, very similar to what happens when my 16 month old drags her favorite Teddy bear over my keyboard; and now that I read what you've written, I agree that it's best not to let 16 month olds and their Teddy bears write firmware and drivers.
Most firmware is in C due to maintainability and low performance hit. Compilers are so good nowadays that it is hardly any worse than hand written assembly.
Thanks for clearing that up. I had heard here and there that a good compiler can rival the performance of assembly, but I ain't gots the learnin' to knows any better. All of my language knowledge is in interpreted languages, so anything at that low a level is akin to black magic to me. I have looked at Assembly code before, and it looks very, very similar to what happens when my 16 month old drags her favorite Teddy bear over my keyboard; and now that I read what you've written, I agree that it's best not to let 16 month olds and their Teddy bears write firmware and drivers.
- Status
Similar threads
