Sdhc - How Fast Is Fast Enough?

[Grench]

OK, so there are 102 or so burn in tested motherboards sitting in a rack.

I'm assuming that someone also has a jig to power one up.

Would there be any chance for that someone to run some simple benchmarks of class 2, class 4, class 6 and -maybe- a class 10 SDHC card?

From what I can tell on Tom's Hardware...

Average read rates:
Class 4 ~ 15MB/s
Class 6 ~ 19.6MB/s
Class 10 ~ ? (According to Sandisk they're 30MB/s, but I haven't seen any real marks on it yet)

The basic question - is the interface on the Pandora capable of keeping up with a Class 10 card? A class 6 card? Where is the point of diminishing returns?

Yes, the class 10 cards are $210 a 32GB pop - twice the price of a 32GB class 4. However, they -might- be 50% faster too. That -might- be worth it if the device can take advantage of it.

So, how fast should we bother to go on SDHC media?

 

[zhasha]

So, how fast should we bother to go on SDHC media?

There's another thread on this. To paraphrase (if I recall correctly): the bus itself can't exceed class 4 speeds.

 

[Grench]

So, how fast should we bother to go on SDHC media?

There's another thread on this. To paraphrase (if I recall correctly): the bus itself can't exceed class 4 speeds.

I saw that - and from what I could tell everyone was simply speculating since there were no real tests. I didn't post to the old thread since I've gotten chewed enough for doing that before.

So, with the 105 tested = 102 passed, sitting and waiting for cases to put them into... How about a benchmark run against some SDHC cards to find out? I'm not asking for a conclusive run all the cards produced benchmark extravaganza - just one card of each available speed class so we can see where it tops out.

 

[zhasha]

The whole "class" business is pretty shady. Some measure read speed, some measure write speed. At the end of the day, the only thing they all have in common is that they all agree on the number corresponding to the lowest speed measured in MB/s, but whether it's reading speed, writing speed, or invisible pink unicorns, who knows. In any case, a Class 4 card is probably all you need unless you have a faster writer somewhere and want to transfer a quadrillion movies onto it.

 

[may88]

The class system specifies the minimum speed of that class. Class 4 is faster than 4Mb/s and 6 >6Mb/s. Not so quick.

I have a class 6 8Gb Transcend that is about 8Mb/s write and a class 4 16Gb that writes at 15mb/s. Hardly scientific!

A lots has been said here

use this to measure the speed of your card

and look here to find a utility that will reformat it for optimum performance.

edit: spacing after the links.

 

[WizardStan]

Yes, the class 10 cards are $210 a 32GB pop - twice the price of a 32GB class 4. However, they -might- be 50% faster too. That -might- be worth it if the device can take advantage of it.

What could you possibly be doing that would require that kind of speed? :blink:

So, how fast should we bother to go on SDHC media?

It really depends on what you want to do with it. If we assume that the minimum read speed = the minimum write speed (a reasonable assumption, I think) then a class 2 is fast enough to play back a 4 gig video file in 45 minutes. If you have a 4 gig file that needs to be played back in less than 45 minutes, you have bigger problems than read/write speed.
SNES games will load in about a second (or 2) at most on a class 2. PSX only had a 2x CDROM drive, so a class 2 outstrips it by quite a bit.
Of course the downside is that, while you can play back a 4 gig video file in 45 minutes, it can still take up to 45 minutes to write that same 4 gig file in the first place. If you try to use the SD card as a swap, you're in for a painful surprise, and if you're booting the OS off such an SD card you're likely to see noticeable pauses as you start up applications. A class 4 or class 6, however, should have no problems starting applications without a visible pause, or so says my opinion. Maybe you would notice a quarter of a second where I would not.
And remember, this's if the minimum read speed is the minimum write speed is the constant speed. 2 mbps may be the minimum for a class 2, but they actually tend to be about 3 or 4 times faster than that on average, at which point that same 4 gig video file plays back in about 15 minutes.
So yeah, unless you're doing something really, really, REALLY read/write intensive, I wouldn't suggest needing more than a class 4.

 

[juniorm33]

My view is pretty simple:
For Pandora general use Class 4 should be fine.
For copying movies a lot etc say from pc to SD faster the better.
So just depends on how you intend to use it.
Loads of roms that wont ever be moved from the card etc class 4 is fine.

 

[may88]

How do you know the unicorns are pink?

I didn't. Do they eat shrimp?

 

[Grench]

How do you know the unicorns are pink?

I didn't. Do they eat shrimp?

"Speeds all a question of money. How fast ya wanna go?" - a mechanic in one of the MadMax movies.

And we're back around to speculation and the 'who needs speed when all ROMs are tiny' bits. Who needs speed when a movie only requires 4MB/sec?

I may not ever even load a ROM to this - I know you're all shocked and surprised.

I may, however, want to boot Ubuntu from the SDHC card, use the SDHC for swap (yeah, I know - wear), load OpenOffice applications and Firefox from it, install the entire contents of Wikipedia for searching through, and may even want to put a PostgreSQL database on it. This kind of general computing is potentially going to push the data throughput to whatever point the OpenPandora device or the SDHC card limits it to.

If the OpenPandora can max out the class 10 cards - I might shell out the extra $100 for the 50% throughput increase. If not, then I have my eye on the Kingston 32GB class 4 cards - I already have 1 of these in my laptop and it's worked well for me.

Back to the original question - which one is the limiting factor for reading data? Does the OpenPandora unit itself have a -real world- maximum throughput that is more or less than the fastest SDHC cards? If it is less, how much less so we can all save a bit of $$ on which cards we buy.

The final boards now exist. It is theoretically possible to answer this question. If there isn't time to, I understand and we can all benchmark stuff later. If there is time to do a quick test/check while we're waiting on cases, then I can keep an eye out for holiday sales bargains on whatever is going to run the fastest in it.

 

[juniorm33]

I'm not surprised you may never load a rom on it and my example about movie is, that your likely to put a movie or series on a card deleted it then put another on, etc. Not the Bandwidth that the movie needs to play.
It all depends on how you intend to use the card.
For general use class 4 should be enough
If your going to be moving a lot of files often and don't want to wait around go for a faster card.
If your going to be booting up from the card then faster wouldn't hurt any.
It's about what you can afford, and what your going to be doing.
The Pandora shouldn't have any problems using the speed to it's fullest. AFAIK

 

[Paradox]

I bought 2x SanDisk Ultra II class 4 32GB cards for my pandora. I'm thinking I should have just sprung for a mini portable USB drive instead.

 

[juniorm33]

Think it's time for me to see what I can pick up over here in the uk. I'll probably go 16GB

 

[frankjr]

I wonder how most external hard drives would compare, since I'm hoping to use the Pandora plus a portable USB-powered external hard drive as a lighter portable media center (I usually go to other people's houses with my laptop and play shows/movies through svideo.)

 

[WizardStan]

I may, however, want to boot Ubuntu from the SDHC card, use the SDHC for swap (yeah, I know - wear)

Some people worry about wear. I don't even get that far. The fastest SD card (at around 30 mbps) is slower than the average hard drive (up around 40 or 50 mbps depending on which chart you read). Unless you are shelling out HUGE amounts of money for the best of the best of the best, and the Pandora's bus handle it (which I'm pretty sure it can't), swap will be SLOW.

 

[Mr B]

OK, so there are 102 or so burn in tested motherboards sitting in a rack.

I'm assuming that someone also has a jig to power one up.

Not sure what your talking about. The "test" boards aren't stacked up in a rack, and set to run. They are shipped of to devs, who've (hopefully) been using them to run what ever gizmos they have been developing, so that when the darn thing finally ends up in the hands of there buyers, there will be some software for it. How they selected what devs would get, and wouldn't get a board? No idea, i would simply have looked on the developers that used to do the old GP2X stuff, and ask if they were up for it.

But thats just what i would do...
B!

 

[Kramy]

I may, however, want to boot Ubuntu from the SDHC card, use the SDHC for swap (yeah, I know - wear)

Some people worry about wear. I don't even get that far. The fastest SD card (at around 30 mbps) is slower than the average hard drive (up around 40 or 50 mbps depending on which chart you read). Unless you are shelling out HUGE amounts of money for the best of the best of the best, and the Pandora's bus handle it (which I'm pretty sure it can't), swap will be SLOW.

mbps == mbit

You're looking for:

MB/sec

The fastest SD cards are almost 300mbit, and the fastest HDDs are almost 1.8gbit. The fastest SSDs are almost 12gbit.

Keep in mind that bits used may not correspond directly to MB/sec transfer speeds, because of error checking bits and stuff.

 

[WizardStan]

The fastest SD cards are almost 300mbit, and the fastest HDDs are almost 1.8gbit. The fastest SSDs are almost 12gbit.

Keep in mind that bits used may not correspond directly to MB/sec transfer speeds, because of error checking bits and stuff.

Aye, the numbers may have been off, but the point still stands: SD is significantly slower than traditional hard drives, suggesting it is probably too slow for swap to be truly effective.

 

[Grench]

Most of you that replied to this thread - you really have no clue what you're talking about. You're real adamant about your opinions, I'll hand you that. Still, no real world experience then. Amazing.

Alrighty then - lets lay out some facts. I'm not going to bother quoting from the mis-information in the posts above. Not worth it.

The 102 successfully tested OpenPandora motherboards known to be in existence were last seen stacked in a pair of racks.

The worlds fastest consumer SSD transfers data at 260MB/s Megabytes per second. Yes there are some industrial raid ones that are faster - but none of us could afford them.
Engadget - OCZ's Colossus

A standard notebook HDD - The Samsung Spinpoint M6 500GB drive in my notebook computer transfers data at 57MB/s.
Tom's Hardware 2.5" drive benchmarks

The fastest desktop HDD - The Western Digital VelociRaptor clocks in at 102MB/s. Yes, you could RAID 0 4 of them an reach about 350MB/s.

The class 10 SDHC cards I'm referring to transfer data at 30MB/s. Yes, that is 1/2 the speed of the hotshot high end drive in my notebook computer.
Sandisk 32GB Extreme III SDHC card

Follow the Tom's hardware 2.5" notebook HDD list down to where the 3 year old 160GB drives are at and - guess what? 2 years ago the 'normal' HDD had slower average read times than these new hotshot SDHC cards.
Tom's hardware 2.5" HDD listings

These new SDHC cards are plenty fast for swap - IF the OpenPandora console can take the data rates.

So, now that we have established that most of the above posters have no clue what the hell they are talking about. Can we stop speculating and see if just maybe someone with access to the actual motherboards might be interested in answering a question before it gets buried 5 pages deep in crap replies?

Is there the will or the time for anyone to run a set of average read benchmarks on an actual Pandora motherboard to fill us in on if it can handle the throughput from these new class 10 cards? I don't think I'm alone in wanting to get past speculation from the clueless and down to some real facts.

Thank you,

Grench

 

[WizardStan]

2 years ago the 'normal' HDD had slower average read times than these new hotshot SDHC cards.

These new SDHC cards are plenty fast for swap - IF the OpenPandora console can take the data rates.

Yeah, the new class 10s possibly. As I said

Unless you are shelling out HUGE amounts of money for the best of the best of the best

Class 4s (and lower end 6s) still come in below what we had even 3 years ago. Even 6 years ago, hard drives were pushing 40 MB/s. I stand by my statement: by the standards of the last 6 years at least, all but the latest and most expensive cards will be too slow for effective swap. That's not to say it can't be done, and won't work well enough if your memory manager is clever, it'll just introduce noticeable lag from time to time.
I don't understand why you seem to be getting so angry about this. You asked a question, and in the absence of one of the official developers, we did our best to answer your question.
Why does this even matter anyway? Once you've got your Pandora, you can perform all the tests you want. If you didn't order one, then once enough people have theirs, I'm sure someone will find the time to perform the tests and give you feedback you crave so you can decide whether it's worth it or not.
Seriously though, if the ability to swap memory at a reasonable speed is the deciding factor for you, I suggest you reexamine your priorities.

 

[kaprikawn]

Is there the will or the time for anyone to run a set of average read benchmarks on an actual Pandora motherboard to fill us in on if it can handle the throughput from these new class 10 cards? I don't think I'm alone in wanting to get past speculation from the clueless and down to some real facts.

It's going to be released soon. Why not just wait and do the benchmarks yourself instead of joining the hoard of people constantly going "Can I have a video of this...?", "Can you test that...?"? I'm sure the devs and OP team have got enough to do without constant, and in my opinion unreasonable, demands. They do a great job of keeping us informed of everything they can, but they're not Tom's Hardware, they clearly don't have time to be testing and benchmarking everything. Give them a break.