Surprisingly cheap deal on Intel 330 SSDs - great price - Page 2

I own a Samsung. And a Crucial. And an Intel.

While performance benchmarkings are neat to look at there is almost no correlation to HTPC use. I can't tell a bit of difference between any of my SSDs and don't even care what their benchmarks are.

I want whatever is the most reliable because of this.

Just pointing it out since the subject came up...

It's pure speculation but with typical desktop or HTPC uses, I bet you wouldn't see a significant real world noticeable difference one way or another. Any differences definitely wouldn't be to the point that you'd notice switching from a spinning drive to solid state.

If you look at the overall performance of current gen drives, they form a bell curve. A few excellent drives (which come with premium prices), most drives lumped together in the middle, and a few crappy drives that should be avoided. IOps, MB/sec...all the relevant statistics hold this way.

For me, price, capacity, and reliability (in some order) are all more important than having the absolute fastest drive. However if I was looking for say a database server for a corporate environment where money wasn't the critical factor but raw performance was, then your original question may warrant further investigation.

Good posts.

That's why when some members on this forum constantly tout SSD X has better performance than SSD Y it drives me crazy.

To be honest you all the SSD's have it ironed out pretty good. Be it Sandforce, Micron,etc. There is no reason for the prices they are selling for today not to have one in your build. The same people who said stay with kerosene lamps, that light bulb is dangerous is the same group.

LOL, I have about half of my house on LED lamps and probably ten SSDs to my name but still use a HDD in my daily driver.

If you look at real data, it is not a clear cut advantage to use SSD. These are my own data or from real owners:

Uh, to be honest I don't even see how you could look at that data and not see a clear cut advantage from the benchmarks... Unless you're trying to compare a crappy low end Sandisk SSD that nobody should be buying (or even a years old Intel SSD, which blows the HDD away, despite the fact that practically any SSD you buy now is going to be even faster than the X25M) to a high end HDD.

There are three types of benchmarks because in common operation all three are used. SSD show fast 4K write performance because it is operating to the controller DRAM buffer, until the DRAM buffer is full. Fast 4K read is real, no buffer. There are not a clear cut advantage in large file size random and sequential operation. This data was actually taken almost two years ago.

What is not shown in this benchmark data is erase operation for SSD. That is dead slow. Erase happens when the disk is full of dirty free space. "Dirty" means free space is not all "1111"s. NAND cannot write 1 over 0s. When there are 0s in free space, that free space must be erased. It is not noticeable when the SSD is new. As it ages, erase operation will be more and more.

Windows7 has a background erase function called TRIM. When the OS is idling, it issues a ready signal to the SSD. Which then starts the garbage collection operation (erase). It is not uncommon for Windows7 not to recognize the device is in fact a SSD or that it supports TRIM.

I am using that 250GB Seagate in my daily driver, partially because I needed more space than economically allows in SSD. Meanwhile I have lots of 32GB and 64GB SSDs.

I did go for that 180GB intel deal.

So what if sequential read/write is similar to hard drives? With how expensive SSDs are, it's unlikely you'll be using them for large file storage (e.g. Blu-ray/DVD rips) which is when fast sequential read/write is useful. The big advantage to SSDs are the near instantaneous access times and much faster random read/write. On an SSD, you could load several largish apps simultaneously without it skipping a beat. On an HDD, loading several apps at the same time usually results in much slowdown. Ever minimized Firefox, left it running in the background for a while and tried opening it again? On an HDD, it seems to take forever to reload. No such problems on an SSD.

Also, the DRAM buffer isn't that big. If the benchmark was only writing to DRAM, then the random 4k writes would have been much faster than what's on your table.

I reckon your issues with SSD performance not being far from HDDs stems more from having crappy quality SSDs (most 32GB SSDs I've seen have crappy controllers).

I don't understand what you are saying. Random 4K write is much faster than HDD. That's because HDD has seek time in random reads and writes. Sequential or large file random is comparable because there are little seek time.

I will run some bench mark data once that 180GB arrives and may be on the 128GB SSD in a work computer.

Yeah, random 4K write of NAND-based SSD is much faster than HDD. What I'm saying is if all it's doing is writing to DRAM instead of NAND, then it should be even faster. Another thing, some SSD controllers don't even make use of a DRAM write buffer (think I read that on AnandTech).

Without write buffer? Find and read that article again.

Actually my work computer (i5) has a 128GB Samsung SSD and home computer (C2D) has a 250GB Seagate HDD. Not much perceived difference in how fast they work.

Of course I never turn off the home computer and the browser has may be 20 panes that is semi-permanently on. It only uses sleep mode and writes eveything to DRAM.

The work computer with SSD is never turned off either (unless it crashes).

You mean reading someone else's data is meaningful and real life?

+1