Forum Vs Naplam - Fasted real world storage solution

[NapalmV5]

heres the final vid.. reboot/reboot script

http://www.youtube.com/watch?v=9buLs4gIotU

not the fastest bootup/reboot but typical


if you guys cant openup your minds and let a bit of illumination in.. go ahead still believe all the useless benchmarks you want.. fine by me

[Boogerlad]

It's not that we don't believe you. I do, but we just want more evidence. I think what they want is for you to reboot, load up cod4 and cod 5 and then the 100 app .bat, because they want to make sure it's not massive caching.

[thebluemeanie1]

While you're at it throw a stop watch in the video so no one complains about the video being fast forwarded.

I have a basic two HDD RAID 0. It isn't incredible or anything but considering your setup cost 10 times as much as mine the speed increase is certainly believable.

[Spoiler]

I'm satisfied with your movie Napalm. Very nice XP boot! You have a pretty good system.

I'm very happy that I got to see your Areca card boot too. I always wondered how long that thing took to initialize. I couldn't tell from the video though: was it the Areca that was taking the most time? Or was it the motherboard posting? If it was the Areca, then it doesn't appear to be that much faster than the Adaptec (only in reference to initialization).

Oh, and one last question:
When are you moving to Vista or Win 7?

[XS Janus]

Wow, nice boot 50sec out of which 20sec go to Areca intializing.
What I would like to see "timed" on these SSDs is that app loading script and COD4,5 load times on ICH vs. Areca!

I know it is probably do tiresome to do, but if you ever do please post results.
I think that would be the final thing that would convince to buy Areca and not go just ICH!

[F.E.A.R.]

I'm satisfied with your movie Napalm. Very nice XP boot! You have a pretty good system.

I'm very happy that I got to see your Areca card boot too. I always wondered how long that thing took to initialize. I couldn't tell from the video though: was it the Areca that was taking the most time? Or was it the motherboard posting? If it was the Areca, then it doesn't appear to be that much faster than the Adaptec (only in reference to initialization).

Oh, and one last question:
When are you moving to Vista or Win 7?

The controller checks all ports. This costs the most time.
Ok, every MB initialise drives faster, but a controller gives more safety

[noobieocer]

hang on reading through this i getting more confused.... so you have ram for a raid controller? and why are you using that card if the southbridge raid is faster?

[Mic]

that boot time was just crazy! and I thought my old Iram Raid 0 with xp was pretty fast to boot.

[F.E.A.R.]

hang on reading through this i getting more confused.... so you have ram for a raid controller? and why are you using that card if the southbridge raid is faster?

Southbridge ist faster? Never!!!

[noobieocer]

Southbridge ist faster? Never!!!

http://www.xtremesystems.org/forums/...d.php?t=207557

o oops i just looked at burst speed lol

[SoulsCollective]

hang on reading through this i getting more confused.... so you have ram for a raid controller? and why are you using that card if the southbridge raid is faster?

If you're referring to FEAR's post above, note that the term used was "initialised faster" - not faster per se. We've pretty much proven in this thread that in real-world usage scenarios, not just benchmark numbers, the Areca card puts out much greater performance than ICH RAID arrays. However, as mentioned, hardware RAID controllers take a long, long time to boot up, because they scan all ports for devices, recognise existing arrays, etc etc.

As for "RAM for a raid controller", sure. Any decent hardware RAID controller will feature on-card cache. The Areca cards just allow you to use a standard DIMM module instead of having the cache RAM soldered to the card, so you can upgrade the cache amount should you choose to.

[Rhys]

We've pretty much proven in this thread that in real-world usage scenarios, not just benchmark numbers, the Areca card puts out much greater performance than ICH RAID arrays.

That's the thing, it hasn't! We are relying on what someone is telling us instead of someone showing us!

How many times have I requested to film the whole thing rather than just a portion that leaves room for speculation of the process being run from caching (Camera should also be handheld also).
I appreciate the work Napalm's put in to 're-educate' some of us, but I can only take the results with a grain of salt because of this.

i.e. if someone said ICH10R was faster than Areca I wouldn't even dream of arguing any different because I still don't have any credible evidence to suggest otherwise.

My request is not anymore difficult than what has been demonstrated already, hence why at this stage the current results are less credible than when they were first published.

You filmed the bootup Napalm, it was great, I have no argument with it at all!
Why can't you do that and film the bat file and COD4/5? and possible demonstrate your camera holding abilities?

If you can do that then I bet 99% of us will change our tune entirely and won't be able to argue with the results.

[One_Hertz]

Why don't you guys believe him? It definitely is odd that he won't post a vid showing a cold boot in front of the tests but the COD4 loading time makes some sense. My guess is that it is well optimized and the requests are mainly sequential reads. Most games are not like this, but some are (I think crysis is like this too). My 2x X25-E load that level in ~2.8 seconds.

20 seconds @ 70MB/s read (7200rpm storage HDD)
2.8 seconds @ 470MB/s read (2x X25-E, very degraded)
1.5 seconds @ 750MB/s read (Napalm's setup)

Makes sense as the numbers roughly match up and there is always these small random differences system to system, especially when getting into loading times that are in the few second range and measuring with a stop watch. Has nothing to do with areca. Any raid card that can do 750mb/s sequentially should be able to do this for COD4. A third X25-E would get me to about 1.9-2.0 seconds and that is where ICH would max out.

By the way, if any of you guys are testing this without at least a 4Ghz Intel and some quick ram then I don't think you'll like what you will see. Your CPU is holding you back and you probably won't load that fast regardless of your storage speed. I did a quick test @ 3Ghz and it took like 5 seconds (my guess, I did not measure with a stop watch) to load versus the 2.8s @ 4.7ghz.

COD4 isn't a good test... The 100app test is much better, however it is also much more difficult to replicate on different systems. Someone make a list of 100 free apps that we could DL. These apps can't include anything that scans your hardware config (like 3dmarks) because that has nothing to do with speed and will just take different amounts of time depending on your hardware.

[SoulsCollective]

We've seen a camcord of the reboot time. That can't be cached. We've seen the 100-app bat run. That could be cached, but really, at this point, I'm prepared to take Napalm's word for it.

[F.E.A.R.]

http://www.xtremesystems.org/forums/...d.php?t=207557

o oops i just looked at burst speed lol

But you can´t use this theoretical value in practise

~ 1600MB/s. with cache and ~ 800MB/s. without cache

The Areca cards just allow you to use a standard DIMM module instead of having the cache RAM soldered to the card, so you can upgrade the cache amount should you choose to.

The 1231/1261/1280ML have no soldered write-cache.
You must remove the original ram-module (256GB) before using other modules

[noobieocer]

omg i just notice this thing is expensive! :O

http://www.scan.co.uk/Products/Areca...-plus-Ethernet

[Jor3lBR]

omg i just notice this thing is expensive! :O

http://www.scan.co.uk/Products/Areca...-plus-Ethernet

Yeah they are! That's why I would like to see comparisons between the new MegaRAID SAS 9260-4i ($328) and these Areca's ($800) to make the best decision

[F.E.A.R.]

omg i just notice this thing is expensive! :O

http://www.scan.co.uk/Products/Areca...-plus-Ethernet

Hm, it´s really expensive in uk.
In GER you pay 600€ or £510

[Rhys]

Your CPU is holding you back and you probably won't load that fast regardless of your storage speed. I did a quick test @ 3Ghz and it took like 5 seconds (my guess, I did not measure with a stop watch) to load versus the 2.8s @ 4.7ghz.

Makes sense, I get 5 seconds with 3x Vertex & 1x x25-m.

[Spoiler]

By the way, if any of you guys are testing this without at least a 4Ghz Intel and some quick ram then I don't think you'll like what you will see. Your CPU is holding you back and you probably won't load that fast regardless of your storage speed. I did a quick test @ 3Ghz and it took like 5 seconds (my guess, I did not measure with a stop watch) to load versus the 2.8s @ 4.7ghz.

Hmm, I will have to experiment with that on my systems.

COD4 isn't a good test... The 100app test is much better, however it is also much more difficult to replicate on different systems. Someone make a list of 100 free apps that we could DL. These apps can't include anything that scans your hardware config (like 3dmarks) because that has nothing to do with speed and will just take different amounts of time depending on your hardware.

I thought the same thing too. It makes it hard for anywone else to verify/test.

[yngndrw]

It's not that we don't believe you. I do, but we just want more evidence.

Hang on a minute, you're contradicting yourself here.

This is getting ridiculous, it's up to you if you either take his word for it or if you don't believe it, but don't waste his time with your petty arguments.

[Boogerlad]

I believe him. Others are the ones that want more info.

[Griff805]

What does it matter anyway- it's all on a stripped down computer. What does Napalm have installed on the machine he's doing all these tests on? What kind of functionality in Windows is he missing because of his stripped OS?

[F.E.A.R.]

Maybe he wants to challenge me

flash-based SSDs vs. DRAM-based SSDs

[karbonkid]

What does it matter anyway- it's all on a stripped down computer. What does Napalm have installed on the machine he's doing all these tests on? What kind of functionality in Windows is he missing because of his stripped OS?

It seems to me he was able to run 100 perfectly usable apps with no loss of 'functionality' Play some pretty intensive games, massively linked apps like OpenOffice, run a lot of Windows' built-in utilities, etc. Stripped down to the point of just being a benchmark beater is very, very different to just being well-optimized.