Jump to content

SSD vs. HDD


Mr.D.
 Share

30 posts in this topic

Recommended Posts

He did use an example having another person copy the files off of his computer not just him alone doing a big file transfer but a possible instance where he could be watching a movie on an HTPC of some sort while having that buddy copy another movie.

Again the sole purpose is not to lower boot times but rather IO activities such as launching apps just to name one.

 

So, unless this person is connected to his LAN, the bottleneck would be the ISP connection, not the HDD transfer rate. And while I guess I could see how somebody bringing his computer over to your house and then downloading a file off the HTPC while watching a movie on that same HTPC would not be outside of the realm of possibility -- However, I'm not gonna pay a premium for my friend to get his files off my machine 15-30% faster. He can just wait. OR, he could buy an SSD for me!

 

Again, my point is that boot time is the ONLY time I can see this being of any appreciable use. I wrote earlier about how when I launch an app, it stays open until I am done with it. I don't open/close/repeat - even on my general purpose laptop.

 

And nobody has answered my challenge that I made in post #19. I saw one stat post about RAID sets on external thunderbolt enclosures.

 

Gauntlet thrown...

Link to comment
Share on other sites

Before responding to your post #19 challenge, lets decide what really different types of things there are that we ask our computers to do, and feel free to add on. We: Boot it into OS, launch apps to use, use the launched apps, save-move-delete files and documents, do internet stuff like browse and watch videos, play games, download stuff, play media such as music and movies, run other OS's virtually, shut down, sleep, resume from sleep. Out of those things, the internet related ones shouldn't be affected (thank god for that fast broadband). Games mostly unaffected (good thing most people have fast CPU's and GPU's these days). Watching movies and listening to music should be the same unless it's 1080p sequential 3d at 48 frames per second/120hz or something. As for the already running apps, it will depend on what other things on the list they are also doing, an app like CCC, Disk Utility, Text Edit, The Unarchiver, MS Office, Mail, iMovie, Logic, Final Cut, or anything that needs to access or save documents after being launched-regardless of file size(it's all about the random read speed)-will benefit, where other apps like Calculator, Mine Sweeper, ColorSync Utility, Facetime or Chess may not. The rest of the list should all see a speed boost. But the same "it's expensive and not worth the price" argument would have applied to other things as well. I'm sure when SATA drives came out and were more expensive than IDE drives, people said "why do I need a faster SATA HD just to boot faster?", and other's responded by saying there was more to a HD than boot times, and eventually as the prices came down and most mobo's had SATA ports, those people finally converted and never looked back. Same for USB 2.0, 3.0. I don't use my thumb drive very often, but if it was USB 1.1 it would be in the trash. There isn't any one magic thing that will instantly make a PC faster, it is the sum of it's parts, yet is also always hindered by it's weakest link. whatever that may be, even MicroSoft knows this. According to MS and their "WEI", your experience will never exceed 5.9 :) , luckily they also say this means you'll still be able to run all the advanced features included in Windows Vista :) Now I don't really care about booting, shutting down or sleeping cause I don't do any of them unless I have to, laptop hasn't rebooted in a few months, desktop had to for a Windows 8 install the other day (don't waste your time on that one, that {censored} sucks!), but otherwise I'd leave it running for 2 years straight if it were feasible.

 

Because drive manufactures like to flash impressive read/write speeds to help sell drives, and boot times and app launch times are the only things that take enough seconds to complete where showing a 25% speed boost gives a result in seconds, which people can understand as a length of time and not just an abstract idea. It's caused people to get hung up on these aspects of a SSD and not look at what it really does so much better than a traditional HD. Take out booting and app launching, and large file transfers which are all sequential reads/writes, and your left with a whole bunch of little random reads and writes scattered throughout your day and your work. Compare the time difference of the accessing and reading of a few of these and the difference may be measurable only in nano seconds, which doesn't sound very impressive in an advertisement. But a 7,200 RPM sata 2 drive averages 75 to 100 IOPS (Input/Output Operations Per Second) and current sata 3 SSD's average 60,000 to 100,000. This is not just a small difference, it's 914 times faster. 914 times!! That's like going from living in a cave to a moon base. That speed applies to every little scrap that gets accessed on your hard drive, and thus the overall user experience is just noticeably faster. So much in the same way upgrading to a CPU that's got a faster clock speed may only give you a measurable comparison while doing specific tasks like encoding video, it would also be hard to argue that it's not faster overall. There may be only a few areas you can point to your stop watch and say "see it completed XYZ in only 4 seconds, I knew this Ivy Bridge must be a little bit faster than my old Pentium 2", but without having to do that, you know just by using it through out the day that it's a hell of a lot faster. It's because of this that benchmarks exist, essentially cramming a whole days worth of work into a small amount of time to provide an inflated result with lengths of time long enough for us to comprehend.

 

So, for the test, I'll need to make a new clone just prior to running it so OS version, installed apps and updates are 100% identical. So first would need a list of tasks for various apps to perform after they've been launched, and have them execute them via apple script so it can be repeated exactly on each drive without any human interaction to skew the results, and have the routines be long and complex enough to consume a measurable amount of time. So before going through the giant PITA that will be, please be sure it's not just illusory superiority that's clouding your opinion and you truly believe it.

 

vQK1ZuUHBEO3PJKp_y5G0g2.jpg

Link to comment
Share on other sites

With regards to SATA vs IDE... Late model IDE drives were ata-parallel. Current generation hdd and ssd drives are ata-serial. ATA is a subset of SCSI. they have removed all of the non-hard drive functions from the command set (like tape drive calls, etc) and optom8ised the interface for what is only needed - data lanes for a HDD. Ever notice how close an eSATA and an SAS connector look? Well, that's why. SAS is a very similar optimization of the SCSI set.

 

So comparing IDE vs. SATA drives for speed isnt a good comparison as there have been several incremental speed increases within each technology (ie. 66/100/133 ide/pata; SATA 1, 2, 3; Bus speed of PCI adapter cards; RPM of platters, etc...) Even today you have the choice between a 5x00 RPM drive, a 7200 one or a 10,000. Lets not forget the amount of cache on the HDD as well, which will affect the speed. So many things affect the price AND speed here. It has always been a matter of 'is this worth it to me'?

 

Are you also stating that you'd pitch a USB 2.0 thumb drive in favor of a USB 3.0? Do you REALLY need to access the files on that device 40% quicker? What a=on earth are you saving to your USB drive that you can justify pitching something that works fine, and by definition isnt a primary storage medium, in favor of the USB 3.0 flavor of that same thumb drive? Really?

 

I have some HDMI monster cable I'd love to sell you. Its so much better than that $10 {censored} you can buy at wal-mart because... um... well... it says MONSTER on the package!! Never mind that its a digital signal. Never mind that its a short distance. It says MONSTER on the package and it must be better!!

 

I agree that ther eis no one thing that makes a machine faster. But, as I have stated before, I will get more bang for my buck with better CPU, GPU and most of all RAM, than I would with SATA 3, an SSD, or USB 3. If I had a hundred buck to spend, you bet your sweet bippy its going into RAM.

 

A million people? I have posted several links to peoples articles about either the illusionary benefits of SSD or the outright problems of SSDs.

 

And a billion people can be wrong. Hello China!

Link to comment
Share on other sites

I'm aware of the differences in HD's, but trying to keep the post to less than a page, and the point should have been clear regardless of the details. Also I wouldn't pitch a USB 2.0 thumb drive, just a 1.1 version, nor can I see a relation to buying overpriced Monster {censored} HDMI's from Worst Buy, but a more accurate comparison might be made between HDMI cables and VGA or something. So there's no convincing someone who has already made up their mind, but am happy to complete your test so long as the exact methodology and content is agreed on, cause I wouldn't want to go through the trouble just to have the results declared suspect because Netscape wasn't included.

Link to comment
Share on other sites

How dare you not include Netscape!! That is the browser to beat my friend. Especially 1.0. Man that was something...

 

So - seriously. I'll give you my machine, and how I use my machine for parameters - hows that? Since this started off, after all, on how I - ME - couldn't see the benefit, or at least couldn't justify the cost and/or reliability issues associated with a large SSD for my music editing data.

 

Pertinent machine specs:

*ASRock Extreme4-m z77 uATX mainboard

*i5-3570k stock clock of 3.4GHz

*Corsair Ballistix Tracer RAM (2x4GB 1600 8-8-8-24 Cas Latency 8)

*ATi/AMD 6770 GPU (1GB GDDR5 76.8 GB/s memory bandwidth, PCIe 2.1 x16)

*Original to the G5 160GB SATA; 7200 rpm (BOOT) -- I assume this is SATA 1, but was unable to confirm

*Samsung Spinpoint F3 HD103SJ 1 Terabyte (1TB) SATA/300 7200RPM 32MB (DATA)

 

** I am using the SATA 2 ports on my motherboard. the SATA 3 ports remain unused.

 

Pertinent Usage Info:

* Win7 Pro Primary Boot OS

* Audacity 2.0.2 DAW

* Audio files between 10-700 MB (occasionally more, usually not)

* Browser(s): SRWare Iron and Waterfox

 

** Apps are installed to the 160GB drive. All data is on the 1TB drive

 

usually, when I turn the machine on, I either load a browser to download music, or I load Audacity. On occasion, I have had both running, but never using them at the same time. normalizing a 700MB file, or time-shifting it is quite processor intensive, so I wait the 2-5 minutes while its doing its thing. I have been doing this sort of thing since the AMD 486DX4 100 (which to normalize that large of a file would take 16-20 hours- yes HOURS), so either I have developed patience, or I just don't get bored in that 2-5 minute break.

 

I have also used this machine to watch music videos -- usually in MP4 format, occasionally in h.264. I have not edited video... yet. When/if I ever do, I'll be upgrading my GPU.

 

This is the machine that I have my VM of 10.6.8 and 10.8.2 on. Both VMs are stored on the DATA drive. I have viewed MP4 files in the VM, and have tinkered with Audacity in the VM, but if I have work to do, I do it in the Win7 environment. The latency produced my the VM is going to be that bottleneck.

 

With regards to every little scrap of read/write: what is it exactly that is being read and written? when you boot, it loads the needed files into memory. When you launch an app, it loads those needed files into memory. What is the I/O?? Sounds like somebody needs more RAM if the I/O is going all the time like that. And lets say that these files, that - for whatever reason - didn't get loaded into RAM, need to be read and written: if they are so small then how would that ever impact my performance? You said 'the overall user experience is just noticeably faster', yet point to small I/O as the reference for this, something that by nature, inst noticeable. And 'overall user experience' is such a subjective term. As you can see from what my machine is used for, I imagine that my 'user experience' is not standard - nor would most of the folks on this site.

 

i know my primary OS is Win7, so that may negatively impact any type of testing/benchmarking you'd do. I will point out that all the software I referenced is cross-platform. I'd be quite interested to see the difference, on either OS, using those parameters. I have other requirements that prevent me from running OS X as a boot OS of any type, so I can ONLY use it in a VM.

 

With regards to the HDMI cable - I was using that as an example of stuff that is 'certified' for 'high speed' and 'less em interference' on the monster {censored} vs. the el-cheapo cable at wally world with BOTH doing the same thing - difference=price/sucker factor. Apples to Apples. HDMI to VGA is like Apples to Submarines. One is analogue, one is digital. one has HDCP/Ethernet/Power/Audio capabilities, one does not. OK, maybe not Apples to Submarines, but more like Apples to Figs: both are fruit, both are tree borne, one is nasty. And the other is figs. HA!

 

And if you have a USB 1.1 thumb-drive, throw it away - not because its 'slow', but because its small.

usb%20logo.jpg

Link to comment
Share on other sites

 Share

×
×
  • Create New...