Intel 750 Series NVME SSD Review

I gotta disagree with you there, I've experienced what a RAMDrive can do to gaming and it's pretty darn cool to load levels nigh-on instantly :) (I imagine it's game dependent)

Edit: Also wraith, you had a stray 1000 in your first post, it should have been a 1024

I didn't know this until recently but the true value got kilobyte, megabyte, gigabyte is infact in decimal
1000KB = 1MB, 1000MB = 1GB etc

The binary version is called kibibyte, mebibyte and gibibyte

This definition changed in 1998, it's just that no-one told us :P

They did tell us we just did not pay attention.
there was a big argument where hdd mfrs were being criticized for false advertizing capacity, they in turn said they were not lying about the size at all and 1GB was infact 1GB, 1000MB with 1MB being 1000KB, and the issue was always that operating systems used power of 2.
2-4-8-16-32-64-128-256-512-1024 (and so on)
it was quite a thing at the time and made the news. then it seemed to die out but they did tell us about KiB MiB GiB and so on.
Also we still keep the B/b for Byte and bits, with a bit being 1/8th of a BYTE.
although im not so sure that a Kib is still 1/8th of a KiB. "although i am pretty sure it is)

funny thing is all the fuss about the hard disks started because people wanted the hard disk manufacturers to use the same measuring standards that end users would be used too. (power of 2) and a 1TB hard disk should be sold and labeled as 909.5GB which would cause less confusion.
What seems to have happened however is the power of 2 measurement was re defined as a different measurement and hard disk manufacturers were allowed to keep selling 909.5GiB drives as 1TB.
So the confusion remained and extra confusion appeared regarding GiB and GB.

(i still prefer the power of 2 measuring)
 
I would rather the power of 2 measurement. Computers are binary so it only makes sense. Think they use their decimal notation so the can save money by not adding in that little bit more to their drives to get the 1024(MB/GB/etc) that consumers have been sold on for ages. I'm sure it cost next to nothing but when you make a couple ten million they probably save enough to make a difference and not put the effort into it. It makes sense to use the power of 2 base measurement anyways because every OS uses it. If MS and Apple(2 biggest leaders) use the power of 2, so should to drive companies who get their products used by them. Keep it all standardized.
 
+1 to the base 2!!! :) As an old nerd, I much prefer the base 2 standard.

Great video Tom! One question though: you mention the white gold award in the video, but no white gold in the review thread. Why not? Did you change your mind after the video?
 
You must be playing some pretty old games with it then. In modern games the difference between a SATA3 SSD and a RAM Drive is practically nil.

Linus did a video on this a few months back: https://www.youtube.com/watch?v=ywAAHuCshnA

Quite obviously older games would benefit less. They are far less bandwidth intensive on all fronts than modern games. If anything they will become more apparent in speed the way things are going. Though it's not practical and far easier to get a PCIe ssd. Ram drives make a difference to some games as Sub said and is game dependent. Just because Linus posted a video on a few games doesn't make it true.
 
Quite obviously older games would benefit less. They are far less bandwidth intensive on all fronts than modern games. If anything they will become more apparent in speed the way things are going. Though it's not practical and far easier to get a PCIe ssd. Ram drives make a difference to some games as Sub said and is game dependent. Just because Linus posted a video on a few games doesn't make it true.
I have experimented with RAM drives a number of times and my results match Linus' results: there is very little if any benefit in most if not all games.

If you know of any games which do benefit from a RAM Drive, with data to back that up, I'd be interested in seeing it.

But the disk rarely seems to be what's holding up load times these days.
I guess it's decompression on the CPU or something like that which is the bottleneck?
 
Quite obviously older games would benefit less. They are far less bandwidth intensive on all fronts than modern games. If anything they will become more apparent in speed the way things are going. Though it's not practical and far easier to get a PCIe ssd. Ram drives make a difference to some games as Sub said and is game dependent. Just because Linus posted a video on a few games doesn't make it true.

It honestly depends on the streaming model in the game. Do they choose to eat all the RAM available or do they cap it? If they cap it then when they stream you will see an improvement.
 
WOW they know how to charge yes i understand its new tech but really

Intel 750 Series 400GB SSD PCI-Express $609.00AUD
Intel 750 Series 1.2TB SSD PCI-Express $1629.00AUD
 
I have experimented with RAM drives a number of times and my results match Linus' results: there is very little if any benefit in most if not all games.

If you know of any games which do benefit from a RAM Drive, with data to back that up, I'd be interested in seeing it.

But the disk rarely seems to be what's holding up load times these days.
I guess it's decompression on the CPU or something like that which is the bottleneck?

This. I used to load games from SSDs in RAID0 back in the day. Load times versus a single SSD were mostly the same. Games right now require less than the full throughput potential of a single SSD because of their inherent lack of multithreading. Maybe things will change when NVMe and DX12 become a thing and SSDs are way more mainsteam. At least one can hope. But for the foreseeable future Serial ATA SSDs are yet sufficient for everything outside of professional applications.
 
Form Factor

The current specification for the Intel 750 series Add-in Cards is:

Add-in Card (AIC) Form Factor
Half-height, Half-length
 Single slot x4 connector

Will this product be available with exchange bracket to:

o Add-in Card (AIC) Form Factor
Full height, Half-length
 Single slot x4 connector

Please see the image
 

Attachments

  • Ful Hight PCIe 3.0 slot.JPG
    Ful Hight PCIe 3.0 slot.JPG
    70.1 KB · Views: 102
My mistake, I have to stop posting with headaches.

But when I first posted, it was the opposite that's shown in the graphs (unless the guv' changed it): graphs showed write faster than read.

(I explained in my previous post why I found it illogical for writes to be faster than reads. ><)

EDIT: quote of myself :p

I am trying to understand how you guys got Write speeds faster than read speeds.

Write speeds are slower because the values have to be written, then checked for accuracy, before switching to next values.
Read speeds should be higher at every test than write speeds. Yet it doesn't do in the graphs.
Is it a firmware problem?

EDIT 2: In Anvil SSD Benchmark, sequential speeds are respected (2.3 Gb/s read, 1.3 Gb/s write). The 128k, 32k and 4k values are all faster in write vs read.
 
Last edited:
X79 nvme

I was wondering if Tom L could poke Asus and ask then when/if they plan to release a BIOS update for the X79 boards to support NVME?

sorry if off topic
 
Youd have to remind me in a couple weeks

Alright I will :)

Right now my SM951 doesn't even show up in BIOS. (I have the ACHI version)

Not to mention that the last released BIOS for RIVBE came out over a year ago, and is a beta (!)

It kinda feels like they have unintentionally forsaken it.
or maybe it was benign neglect, what do i know.

I like the speed on the drive though.

Ivzxg5J.png
 
Back
Top