Never ending wait for NVMe

[iiiankiii]

Were they identical pairs of SSD's? If you are copying from one model to another you are going to be limited by the respective read and write speed.

Of those 3 games from what I remember Skyrim and counterstrike loaded pretty quickly on an HDD.

Could you try Dragon Age Inquisition?

They're both the same SSD. They're both 1TB crucial m550. Like I said, I'm pretty sure NVMe would smoke an SATA SSD on large file transfer. I don't dispute that at all.

I don't think I can try a game that huge. I'm on a z97 motherboard that is limited to 32GB. I think I can try it on a game that is around 20GB or so.

 

[larryccf]

i'm curious to know why the speed increase i've seen and reported earlier in this thread, isn't supposed to be?

For the record - my system is:
cpu - intel i7-4790
ram - G.skill 1600 Mhz DDR3, 16 GB
OS drive - samsung xp941 256 GB
GPU - asus GTX 750 Ti, 2 gb
mobo - asus z97M-plus

Installing the OS to the xp - I cloned the OS drive (another 840 EVO) to the xp941, then used windows to "repair" the boot files that were either missing or corrupted, so whatever settings etc, should have remained constant from the sata SSD to the xp941 installation.

when i first copy a BD file from the BD disk, i use MakeMKV and copy it to a samsung 840 EVO. That time is consistent with what it was before i installed the xp941, but that copy/write speed is limited by the "read" speed of the BD rewriter.

Then from there, i use Handshake to convert or render it to mp4-H.264, writing it to the "D" partition i created on the samsung xp941.

The handshake program, as most video programs, displays "time remaining" - that displayed time never seems to be accurate or rather keeps adjusting upwards till the program has rendered about 8% - 10% of the file. From that point on, it remains consistent (and fairly accurate). And, for the record, when rendering a file using Handshake, cpu usage shows 99-100%. CPU core temps also confirm that usage (temps run 61 - 66C as reported by RealTempGT).

But from day one of the xp941's installation, the decrease in "time remaining" jumped out at me as it's the first thing i look at, after starting the render.

From one 34 GB video file to another, the actual render time may vary by as much as 10%, but i've assumed that's due to the codec of the original file - as i understand it, the MKV format is simply a "shell" format holding a 1:1 copy of the original file, similiar to an ISO file.

But still, a 34-36 GB file runs 28-35 minutes (and occasionally as high as 41 minutes) versus the 65-75 (or more) minutes writing to a sata SSD. That decrease in time writing to the xp941 seems to correlate consistent to the write speed increase of the xp941 (860 MB/s) vs the write speed of the 840 EVO (490-510 MB/s). I realize that when rendering, reading from the 840 EVO is limited by the sata read speed (550-560 MB/s) but not sure how to reconcile that, ....possibly the data is moved to my ram where it's then rendered and written to the xp941.

And apparently i'm not alone in my observations - back earlier in this thread, another video hobbyist reported similiar speed increase with his M.2 PCIe SSD.

 

[phillyman36]

Because the NVME reads and writes faster wouldn't it speeds things up? Going from a mechanical drive to an ssd I noticed an improvement in the time it takes to get things done when using software like Clown Bd. I have a 750 coming today(other specs are in my signature).

 

[Deders]

And, for the record, when rendering a file using Handshake, cpu usage shows 99-100%. CPU core temps also confirm that usage (temps run 61 - 66C as reported by RealTempGT).

Intreaged by this, If the CPU s at 100% doesn't that imply some encoding is going on? or maybe it is demuxing the copy protection?

 

[Redstorm]

i'm curious to know why the speed increase i've seen and reported earlier in this thread, isn't supposed to be?

OS drive - samsung xp941 256 GB

Because the NVME reads and writes faster wouldn't it speeds things up? Going from a mechanical drive to an ssd I noticed an improvement in the time it takes to get things done when using software like Clown Bd. I have a 750 coming today(other specs are in my signature).

The xp941 is an AHCI M.2 SSD not NVMe. its still using the older protocol.

 

[Redstorm]

That's not a fair statement at all. Besides, copying large files, I hardly notice any difference between a gen 1 SSD (old school vertex 2) and the newer, faster SSD (Sandisk extreme). It just isn't too noticeable. I don't think people are trolling.

Plus, awhile back, I have tested RAM disk against SATA SSD. I noticed no difference in game load time at all. The time was nearly identical. Considering that RAM disk is way, way, way faster than NVMe drives (like 2 or 3 times faster), I don't expect NVMe to load games, and programs noticeably faster than SATA.

Your not comparing apples to apples, your logic is flawed, you state RAM disk is faster than a non volatile SSD, then go on to compare NVMe vs SATA SSDs where it is plainly obvious from the testing is faster by 3 to 4 times, pure bandwidth available alone 3200MB vs 550MB.

I can guarantee you that from power on to game loaded from a NVMe SSD will beat your RAM Drive every day of the week, (as your busy installing your game to your RAM disk every time you switch on.)

Dam it, I got sucked in and feed the troll.

 

[bradly1101]

Your not comparing apples to apples, your logic is flawed, you state RAM disk is faster than a non volatile SSD, then go on to compare NVMe vs SATA SSDs where it is plainly obvious from the testing is faster by 3 to 4 times, pure bandwidth available alone 3200MB vs 550MB.

I can guarantee you that from power on to game loaded from a NVMe SSD will beat your RAM Drive every day of the week, (as your busy installing your game to your RAM disk every time you switch on.)

Dam it, I got sucked in and feed the troll.

It's kinda funny that ALL of his hardware is only marginally better than the next cheapest option, but doesn't consider this drive in that equation. Once you add up all the 'marginal' increases you have a kick ass system.

 

[phillyman36]

Just installed the Intel 750 ssd.

Intel 750 ssd
Clown BD 4:15 to process blu ray
Makemkv :51(seconds) to turn blu ray into mkv file
Handbrake 8:51 encode 3 chapters of blu ray to mkv file

Samsung 840 pro
Clown Bd 7:01 to process blu ray
Make mkv 2:32 to turn blu ray into mkv file
Handbrake 13:40 encode 3 chapters of blu ray to mkv file

After ripping the movie to a folder first each test was done with the movie folder on the drive being tested.

 

[larryccf]

Just installed the Intel 750 ssd.

Intel 750 ssd
.....
Makemkv :51(seconds) to turn blu ray into mkv file

........

dying to know how you got that kind of time on copying a BD disk - i'm right now copying "TAKEN 3" using MakeMKV and time is 29 minutes copying it to the 840 EVO (my "worktable" drive), but it's also one of those that the movie is broken up into a god awful number of files.

but i am glad someone is seeing similiar time improvements, so i don't sit wondering if maybe i'm smoking homerolls when i record these times

 

[phillyman36]

dying to know how you got that kind of time on copying a BD disk - i'm right now copying "TAKEN 3" using MakeMKV and time is 29 minutes copying it to the 840 EVO (my "worktable" drive), but it's also one of those that the movie is broken up into a god awful number of files.

but i am glad someone is seeing similiar time improvements, so i don't sit wondering if maybe i'm smoking homerolls when i record these times

Those times are after I copied the movie to the ssd drive using AnydvdHD. These were just simple tests to see the speed difference between the AHCI ssd vs nvme ssd.

 

[larryccf]

OH- thought you'd copied it from blu-ray drive using MakeMKV in that 51 second time - i was hoping you found some magic lubricant to use on the BD drive

one day i need to look into other software - started with DVDFAB and got so burnt with that junk crapware, when i learned MakeMKV would copy DVDs or BDs, I've just stayed with it

 

[phillyman36]

OH- thought you'd copied it from blu-ray drive using MakeMKV in that 51 second time - i was hoping you found some magic lubricant to use on the BD drive

one day i need to look into other software - started with DVDFAB and got so burnt with that junk crapware, when i learned MakeMKV would copy DVDs or BDs, I've just stayed with it

I use Anydvd hd to rip to a hard drive. From there I will either use handbrake or makemkv.

I'm trying to decide if the speed boost is really worth almost $400 as a secondary drive.

 

[larryccf]

i'd only picked up the xp941 to try a PCIe SSD firsthand, and got lucky, found it for $182

But without it, there's no way i'd be able to do 3-4 videos a night (and some nights i do do that many).

Now it's got me looking at the SM951 NVMe, and moving the xp941 back to the M.2 slot and using it as my "worktable" SSD

I'd still feel the same way if i'd dropped $400 on it, and if i go with the 512 GB vs of the sm951, i will be dropping nearly that amount

 

[Jubassi]

thats amazing really wish the prices were not so high i would get one same day

 

[iiiankiii]

Your not comparing apples to apples, your logic is flawed, you state RAM disk is faster than a non volatile SSD, then go on to compare NVMe vs SATA SSDs where it is plainly obvious from the testing is faster by 3 to 4 times, pure bandwidth available alone 3200MB vs 550MB.

I can guarantee you that from power on to game loaded from a NVMe SSD will beat your RAM Drive every day of the week, (as your busy installing your game to your RAM disk every time you switch on.)

Dam it, I got sucked in and feed the troll.

Here's a benchmark from a RAM DRIVE with DDR3 1600mhz.

Here's the 750:

The last time I check, that's a LOT faster than an NVMe drive? Is it not?

No, you're not getting it. I never said NVMe is slow? Wth? I said IN GAME LOADING SCENERIOES, a RAM disk was barely faster than an SSD drive. Don't you get it? Do you realize that a RAM disk is 3 or 4 times FASTER than your NMVe drive? If I barely noticed the game load time with a RAM disk when compared to my SATA SSD, why would you think an NVMe drive would be faster than a RAM disk? Please explain that logic to me? I am serious. Please help me connect the dots. How is an NVMe faster than a RAM drive? Go ahead. Go explain it to me.

Yes, I get there are real noticeable benefits of NVMe, large file transfer is one of them. However, if you're not doing video editing or moving those files around, please enlighten me where else an NVMe can be beneficial? I really want you to convince me because I'm itching to grab one but need a real solid reason to (excluding video editing stuff because that's not I use my PC for).

Oh, still don't believe me.. Here's the conclusion from a review site:
http://techreport.com/review/28050/intel-750-series-solid-state-drive-reviewed/7

While the 750 Series delivers incredible performance in targeted benchmarks and demanding sequential transfers, it doesn't load big files, applications, or games appreciably faster than older SATA SSDs.

http://www.trustedreviews.com/intel-ssd-750-review-performance-and-verdict-page-2

But the question is whether you really need that speed – and for most users that answer will be an emphatic no. An SSD such as Samsung's 850 Pro will be ample for the majority, and will be fine for the foreseeable future too.

However, if you work in a field that requires intensive computation – video editing or 3D design, for example – the Intel SSD 750 could be the perfect way to take your rig’s performance to the next level.

http://www.pcworld.com/article/2904340/intel-750-series-ssd-review-storage-so-fast-only-the-highest-end-pcs-can-keep-up.html

most client usage is relatively straight-line and won’t take advantage of everything NVMe has to offer. Multi-track audio and video editing, transcoding, and the like may see greater boosts than simply copying a single file.

 

[Redstorm]

The last time I check, that's a LOT faster than an NVMe drive? Is it not?

Lol - please show me where i said a RAM Disk is slow (you really went off the deep end trying to prove your tech prowess their) Everyone here knows a RAM Disk is faster than a SSD by quite a margin today. Your comparing volatile memory to non volatile which have completely different usage scenarios.

I'm going to stop trolling you back now cause you bite too easy, i need more of a challenge.

Can you find, some benchmarks for a RAM drive on a X99 setup, double the number of channels to your above benchmark.

 

[larryccf]

for the sake of the illiterate (namely me), WHAT IS A RAM DISK?

never mind - google was my friend

 

[Deders]

Here's the 750:

The last time I check, that's a LOT faster than an NVMe drive? Is it not?

Would just like to point out that the 4k single depth reads here are no faster than a normal Sata AHCI drive.

 

[larryccf]

nobody??? - i'm still hoping for an answer, especially after being made to feel like i'm smoking home rolled cigarettes in my observations by those commenting how there was no noticeable performance increase in usage over a sata SSD

i'm curious to know why the speed increase i've seen and reported earlier in this thread, isn't supposed to be?

For the record - my system is:
cpu - intel i7-4790
ram - G.skill 1600 Mhz DDR3, 16 GB
OS drive - samsung xp941 256 GB
GPU - asus GTX 750 Ti, 2 gb
mobo - asus z97M-plus

Installing the OS to the xp - I cloned the OS drive (another 840 EVO) to the xp941, then used windows to "repair" the boot files that were either missing or corrupted, so whatever settings etc, should have remained constant from the sata SSD to the xp941 installation.

when i first copy a BD file from the BD disk, i use MakeMKV and copy it to a samsung 840 EVO. That time is consistent with what it was before i installed the xp941, but that copy/write speed is limited by the "read" speed of the BD rewriter.

Then from there, i use Handshake to convert or render it to mp4-H.264, writing it to the "D" partition i created on the samsung xp941.

The handshake program, as most video programs, displays "time remaining" - that displayed time never seems to be accurate or rather keeps adjusting upwards till the program has rendered about 8% - 10% of the file. From that point on, it remains consistent (and fairly accurate). And, for the record, when rendering a file using Handshake, cpu usage shows 99-100%. CPU core temps also confirm that usage (temps run 61 - 66C as reported by RealTempGT).

But from day one of the xp941's installation, the decrease in "time remaining" jumped out at me as it's the first thing i look at, after starting the render.

From one 34 GB video file to another, the actual render time may vary by as much as 10%, but i've assumed that's due to the codec of the original file - as i understand it, the MKV format is simply a "shell" format holding a 1:1 copy of the original file, similiar to an ISO file.

But still, a 34-36 GB file runs 28-35 minutes (and occasionally as high as 41 minutes) versus the 65-75 (or more) minutes writing to a sata SSD. That decrease in time writing to the xp941 seems to correlate consistent to the write speed increase of the xp941 (860 MB/s) vs the write speed of the 840 EVO (490-510 MB/s). I realize that when rendering, reading from the 840 EVO is limited by the sata read speed (550-560 MB/s) but not sure how to reconcile that, ....possibly the data is moved to my ram where it's then rendered and written to the xp941.

And apparently i'm not alone in my observations - back earlier in this thread, another video hobbyist reported similiar speed increase with his M.2 PCIe SSD.

 

[Redstorm]

nobody??? - i'm still hoping for an answer, especially after being made to feel like i'm smoking home rolled cigarettes in my observations by those commenting how there was no noticeable performance increase in usage over a sata SSD

Nice gains their over the obsolete SATA SSD tech their.

Moving to the XP941 with the 4 x PCIe 2.0 interface solves the 550MB observed maximum the SATA3 AHCI SSDs can achieve.

Next step is to move to NVMe or Non-Volatile Memory Host Controller Interface Specification (NVMHCI) to fully unleash the potential of SSD's. Currently underway and eagerly awaited.

 

[utahraptor]

Are any of you running nvme in hardware RAID 0 and if so do you boot from it?

 

[iiiankiii]

Lol - please show me where i said a RAM Disk is slow (you really went off the deep end trying to prove your tech prowess their) Everyone here knows a RAM Disk is faster than a SSD by quite a margin today. Your comparing volatile memory to non volatile which have completely different usage scenarios.

I'm going to stop trolling you back now cause you bite too easy, i need more of a challenge.

Can you find, some benchmarks for a RAM drive on a X99 setup, double the number of channels to your above benchmark.

Look, man. Nobody is trying to troll you or anything. All some of us are saying is that there aren't really any noticeable difference between things like loading up games, programs, booting and most day to day things. I'm just saying, unless you're doing something like workstations, video editing and moving around a lot of files, it might not be worth it.

I get it. People want to best. Why wouldn't we? But, that's like recommending somebody an 8 core Intel CPU for gaming. Yes, there are some instances where you'll see the benefit of an 8 core Intel CPU, but for the most part, you're better off with a 4 core Intel CPU. That's what I'm trying to say. Yes, it be nice to have an 5960x, but, I can do most of the things I do with a 4790k. Why waste money on a 5960x when it's not needed? Would you still recommend it somebody that doesn't do things to take adavantage of the extra cores? No. Don't you agree?

The same applies here. For the most part, a regular SATA SSD is good enough and you'll be hard press to notice a difference by going to a NVMe. I understand the future is NVMe. Progress in tech is nice. When the time comes, I'm sure I'll grab an NVMe (when it's a bit cheaper). Nobody is denying that there aren't benefits to an NVMe SSD. The question is? Besides wanting the best just because, is it worth it if you're not pushing the hardware? It really is an honest question. Like I said before, I am itching to grab an Intel 750 but I am holding back because I'm not sure I would benefit much from it over an SATA SSD.

 

[steve wilson]

Look, man. Nobody is trying to troll you or anything. All some of us are saying is that there aren't really any noticeable difference between things like loading up games, programs, booting and most day to day things. I'm just saying, unless you're doing something like workstations, video editing and moving around a lot of files, it might not be worth it.

I get it. People want to best. Why wouldn't we? But, that's like recommending somebody an 8 core Intel CPU for gaming. Yes, there are some instances where you'll see the benefit of an 8 core Intel CPU, but for the most part, you're better off with a 4 core Intel CPU. That's what I'm trying to say. Yes, it be nice to have an 5960x, but, I can do most of the things I do with a 4790k. Why waste money on a 5960x when it's not needed? Would you still recommend it somebody that doesn't do things to take adavantage of the extra cores? No. Don't you agree?

The same applies here. For the most part, a regular SATA SSD is good enough and you'll be hard press to notice a difference by going to a NVMe. I understand the future is NVMe. Progress in tech is nice. When the time comes, I'm sure I'll grab an NVMe (when it's a bit cheaper). Nobody is denying that there aren't benefits to an NVMe SSD. The question is? Besides wanting the best just because, is it worth it if you're not pushing the hardware? It really is an honest question. Like I said before, I am itching to grab an Intel 750 but I am holding back because I'm not sure I would benefit much from it over an SATA SSD.

Very well put. I still want the Samsung NMVe drive though... logic out the window for me ;p

You probably already know, but just in case, that Intel drive takes a long time to boot up (assuming you are planning to boot from it). It has annoyed some other users on the forums.

 

[StrangerGuy]

Look, man. Nobody is trying to troll you or anything. All some of us are saying is that there aren't really any noticeable difference between things like loading up games, programs, booting and most day to day things. I'm just saying, unless you're doing something like workstations, video editing and moving around a lot of files, it might not be worth it.

I get it. People want to best. Why wouldn't we? But, that's like recommending somebody an 8 core Intel CPU for gaming. Yes, there are some instances where you'll see the benefit of an 8 core Intel CPU, but for the most part, you're better off with a 4 core Intel CPU. That's what I'm trying to say. Yes, it be nice to have an 5960x, but, I can do most of the things I do with a 4790k. Why waste money on a 5960x when it's not needed? Would you still recommend it somebody that doesn't do things to take adavantage of the extra cores? No. Don't you agree?

The same applies here. For the most part, a regular SATA SSD is good enough and you'll be hard press to notice a difference by going to a NVMe. I understand the future is NVMe. Progress in tech is nice. When the time comes, I'm sure I'll grab an NVMe (when it's a bit cheaper). Nobody is denying that there aren't benefits to an NVMe SSD. The question is? Besides wanting the best just because, is it worth it if you're not pushing the hardware? It really is an honest question. Like I said before, I am itching to grab an Intel 750 but I am holding back because I'm not sure I would benefit much from it over an SATA SSD.

i just love how those NVMe whiteknights claiming they are trolled when we have been saying all along that synthetics != general real-world use and we aren't even dimissing niche cases where NVMe can have significant gains. Next time I'm going buy CPUs just by how much memory bandwidth they can push in SisoftSandra.

 

[Redstorm]

I still want the Samsung NMVe driver though... logic out the window for me ;p

ditto