1 |
On 12/02/2020 12:45 PM, Rich Freeman wrote: |
2 |
> On Wed, Dec 2, 2020 at 1:33 PM Frank Steinmetzger <Warp_7@×××.de> wrote: |
3 |
>> |
4 |
>> Am Tue, Dec 01, 2020 at 02:46:56PM -0700 schrieb thelma@×××××××××××.com: |
5 |
>>> Nothing scientific, but I was surprised how fast M.2 disk so decided to |
6 |
>>> time how fast GnuCash will load my accounting, her it is: |
7 |
>>> |
8 |
>>> Box 1.) |
9 |
>>> WD (spinning disk) CPU AMD-8150 (8-core), 16GiB |
10 |
>>> Time to open GnuCash - 23sec. |
11 |
>>> |
12 |
>>> Box 2.) |
13 |
>>> Samsung SSD 850, CPU AMD Ryzen 5 1400 Quad-Core, 16GiB |
14 |
>>> Time to open GnuCash - 15sec. |
15 |
>> |
16 |
>> I can hardly believe that. Does your duration include the entire boot |
17 |
>> process? If so, the times look quite alright, but that’s not what |
18 |
>> description says. |
19 |
>> |
20 |
> |
21 |
> Keep in mind these are different hosts, probably running different |
22 |
> software with different workloads/optimizations, with different CPUs, |
23 |
> differing amounts of RAM, and different storage technologies. |
24 |
> |
25 |
> It isn't particularly controversial to suppose that M.2 (NVMe) is |
26 |
> going to be faster than SATA-based SSD, which is going to be faster |
27 |
> than spinning disks. |
28 |
> |
29 |
> When you want to get to the exact differences you need to test on |
30 |
> configurations that are otherwise identical, and also account for |
31 |
> stuff like caching. |
32 |
> |
33 |
> Note also that M.2 is a form-factor, and you can find SATA-based M.2 |
34 |
> drives which aren't going to perform any better than any other |
35 |
> SATA-based drive. Also, since NVMe is far more capable than SATA/AHCI |
36 |
> it matters even more exactly what NVMe drive you're talking about. |
37 |
> The storage device itself, and the PCIe version can make a difference, |
38 |
> and of course you need a CPU/MB that can actually take advantage of |
39 |
> the drive's full capability. |
40 |
> |
41 |
> You might not need the max performance NVMe is capable of, but it is |
42 |
> something you should be aware of if you want to benchmark it. |
43 |
|
44 |
You are correct, it wasn't any benchmark test. And yes, the old PC are |
45 |
about 10-year old and 3-4 years old; and the new one is just assembled. |
46 |
But I was surprised how fast the GnuCash open the database on a new box, |
47 |
so I time it for a curiosity. |