1
00:00:00,480 --> 00:00:05,839
truth in advertising

2
00:00:03,520 --> 00:00:10,400
whether it's the tact on resort fees from your last vegas vacation or that

3
00:00:08,160 --> 00:00:14,799
photo of the big mac that looks nothing like the burger you'll be eating the

4
00:00:12,240 --> 00:00:19,359
corporations are always finding new misleading ways to market their products

5
00:00:17,119 --> 00:00:24,800
and this is no less true in the tech world especially when it comes to ssds

6
00:00:22,240 --> 00:00:29,519
so sabrint sponsored this video demystifying some of the SSD marketing

7
00:00:27,599 --> 00:00:33,760
that's out there now don't get me wrong when companies claim that ssds are

8
00:00:31,679 --> 00:00:39,040
faster than traditional hard drives they are definitely right about that the

9
00:00:36,399 --> 00:00:44,000
issue is that typically the numbers on the spec sheet or the advertisement from

10
00:00:41,280 --> 00:00:48,000
the manufacturer while very impressive sounding and big

11
00:00:45,520 --> 00:00:53,840
in many cases seem to be designed only to make their drive sound incredibly

12
00:00:50,719 --> 00:00:56,079
fast now in this video we discussed

13
00:00:53,840 --> 00:01:02,559
which specs to pay the most attention to when shopping for an SSD for example an

14
00:00:59,199 --> 00:01:05,119
SSD will be able to read or write large

15
00:01:02,559 --> 00:01:10,799
sequential files something like a space hogging video file more quickly than

16
00:01:07,760 --> 00:01:12,479
lots of small more random bits of data

17
00:01:10,799 --> 00:01:17,920
like you would encounter with frequently used programs like web browsers meaning

18
00:01:15,360 --> 00:01:21,840
that the random performance specs are more relevant to the everyday user than

19
00:01:20,000 --> 00:01:27,520
the sequential ones but here's the rub there isn't any kind of

20
00:01:24,960 --> 00:01:33,520
body or industry association that oversees SSD testing and while ssds work

21
00:01:31,520 --> 00:01:39,040
very differently compared to mechanical hard drives and even compared to each

22
00:01:35,840 --> 00:01:42,079
other in many cases oftentimes the exact

23
00:01:39,040 --> 00:01:43,439
same tests are used on hard drives and

24
00:01:42,079 --> 00:01:47,520
ssds you see most quick and dirty drive

25
00:01:45,439 --> 00:01:52,799
benchmarks measure how long it takes for a drive to accomplish some task say

26
00:01:50,640 --> 00:01:57,920
transferring a large file but this kind of work is a lot more

27
00:01:55,119 --> 00:02:02,560
straightforward for a hard drive because fragmentation aside it can simply

28
00:02:00,240 --> 00:02:09,119
overwrite whatever space on the platter is marked as free ssds on the other hand

29
00:02:05,680 --> 00:02:11,760
often need to erase the data first and

30
00:02:09,119 --> 00:02:15,760
move some existing data before new data can be written to a specific block you

31
00:02:14,000 --> 00:02:20,560
can learn more about that up here but basically it adds more time to your SSD

32
00:02:18,160 --> 00:02:24,480
write operations especially if your SSD is getting full and doesn't have as much

33
00:02:22,400 --> 00:02:28,879
free space to work with and while it might not be difficult to measure how

34
00:02:26,480 --> 00:02:33,760
long this process takes the issue is that in the real world people aren't

35
00:02:31,120 --> 00:02:37,040
writing to mostly empty drives which is what many manufacturers use for their

36
00:02:35,840 --> 00:02:41,920
benchmarks instead as your SSD fills up from daily

37
00:02:39,680 --> 00:02:46,560
use there's more and more performance overhead due to data reshuffling which

38
00:02:44,239 --> 00:02:50,480
can cause real world speeds to drop off dramatically far below what's listed on

39
00:02:49,360 --> 00:02:55,280
the spec sheet then making matters much worse you have

40
00:02:53,200 --> 00:03:00,720
the fact that the number of random reads in particular that an SSD can do this is

41
00:02:57,519 --> 00:03:03,519
measured in iops is often reported while

42
00:03:00,720 --> 00:03:07,440
the SSD isn't also trying to write much data and in the real world there are

43
00:03:05,760 --> 00:03:12,159
many scenarios where you would be doing both reading and writing at the same

44
00:03:09,840 --> 00:03:16,000
time meaning that these read iops that are reported on the sheet can easily end

45
00:03:14,400 --> 00:03:20,000
up inflated throw in the fact that an increasing

46
00:03:17,680 --> 00:03:24,480
number of ssds employ complex data management and caching algorithms to

47
00:03:22,400 --> 00:03:27,360
boost drive performance beyond what should be possible with the underlying

48
00:03:26,080 --> 00:03:32,239
flash memory under certain conditions

49
00:03:29,440 --> 00:03:36,400
and it can be very unclear exactly what kind of data was used to get the numbers

50
00:03:34,400 --> 00:03:41,360
that you see when you go to buy a drive on amazon or newegg so then does that

51
00:03:39,040 --> 00:03:46,400
mean that SSD manufacturers are all evil and going out of their way to lie to you

52
00:03:43,760 --> 00:03:51,360
well actually probably not it's more likely that because proper SSD

53
00:03:48,480 --> 00:03:56,239
benchmarking is way more difficult and time consuming than testing a hard drive

54
00:03:53,760 --> 00:04:01,120
manufacturers simply can't be bothered to develop and run these complicated

55
00:03:59,040 --> 00:04:05,280
benchmarks versus the quick ones that they do on these mostly empty drives

56
00:04:03,280 --> 00:04:09,280
because it not only spits out impressive looking numbers that look great on the

57
00:04:06,640 --> 00:04:13,280
spec sheet but it also saves them time and resources

58
00:04:10,879 --> 00:04:17,280
and the thing is that so far this is only even bitten a handful of

59
00:04:15,280 --> 00:04:22,400
manufacturers in the butt because the truth is that many people who buy an SSD

60
00:04:20,239 --> 00:04:26,320
just want an easy boost in speed and responsiveness compared to their hard

61
00:04:24,320 --> 00:04:30,160
drive and they may not even care about the exact performance numbers

62
00:04:29,040 --> 00:04:34,080
but anyway let's say you really are concerned about

63
00:04:32,160 --> 00:04:37,520
how a drive stacks up before you buy it well for the discerning buyer we've

64
00:04:35,759 --> 00:04:42,160
linked a couple of websites down below that do more comprehensive drive testing

65
00:04:40,080 --> 00:04:46,000
using methods that closely simulate real-world workloads and then drawing

66
00:04:44,400 --> 00:04:50,880
conclusions from that so grab those reading glasses

67
00:04:49,440 --> 00:04:56,240
and get cracking it takes a bit more effort but this knowledge could come in handy for when

68
00:04:54,000 --> 00:05:00,320
you need peak performance for all that super critical work you do

69
00:04:58,479 --> 00:05:05,680
speaking of peak performance check out the rocket NVMe SSD from our sponsor for

70
00:05:02,639 --> 00:05:07,840
today's video sabrant it uses an m.2

71
00:05:05,680 --> 00:05:12,240
interface so that means it's super small not taking up pretty much any room in

72
00:05:09,919 --> 00:05:16,880
your computer and it supports pci express 3.0 gen 4 so it's super fast

73
00:05:14,800 --> 00:05:20,720
whether you're gaming editing videos or completing projects it also features

74
00:05:19,039 --> 00:05:24,240
advanced wear leveling to ensure that it'll be snappy and reliable for a long

75
00:05:22,639 --> 00:05:27,520
time to come and you can check it out today at the link in the video

76
00:05:25,680 --> 00:05:31,199
description so thanks for watching guys like dislike

77
00:05:29,440 --> 00:05:34,720
check out our other videos leave a comment if you have suggestions for

78
00:05:32,720 --> 00:05:39,199
future fastest possibles and don't forget to subscribe right away we do

79
00:05:37,280 --> 00:05:43,960
everything as fast as possible why can't you don't put it off that's how you

80
00:05:40,960 --> 00:05:43,960
forget