1 |
On Fri, Jul 27, 2012 at 8:30 PM, microcai <microcai@×××××××××××××.org> wrote: |
2 |
> 2012/7/26 Michael Mol <mikemol@×××××.com>: |
3 |
>> On Thu, Jul 26, 2012 at 10:18 AM, Евгений Пермяков <permeakra@×××××.com> wrote: |
4 |
>>> On 07/26/2012 05:50 PM, Michael Mol wrote: |
5 |
>>>> |
6 |
>>>> On Thu, Jul 26, 2012 at 9:43 AM, Евгений Пермяков <permeakra@×××××.com> |
7 |
>>>> wrote: |
8 |
>>>>> |
9 |
>>>>> On 07/26/2012 12:05 AM, Philip Webb wrote: |
10 |
>>>>>> |
11 |
>>>>>> I've listed what's available at the local store, |
12 |
>>>>>> which I trust to stock reliable items, tho' I wouldn't ask their advice. |
13 |
>>>>>> |
14 |
>>>>>> All the AMD's are 32 nm , while the Intel recommended by one commenter |
15 |
>>>>>> -- Core i5-3570 4-Core Socket LGA1155, 3.4 Ghz, 6MB L3 Cache, 22 nm -- |
16 |
>>>>>> is 22 nm : it costs CAD 230 & they have 3 in stock, |
17 |
>>>>>> which suggests demand, but not the most popular ( 9 in stock). |
18 |
>>>>>> |
19 |
>>>>>> Isn't 22 nm going to be faster than 32 nm ? |
20 |
>>>>>> |
21 |
>>>>>> In the same price range, AMD offers Bulldozer X8 FX-8150 (125W) |
22 |
>>>>>> 8-Core Socket AM3+, 3.6 GHz, 8Mb Cache, 32 nm ( CAD 220 , 2 in |
23 |
>>>>>> stock). |
24 |
>>>>>> |
25 |
>>>>>> How do you compare cores vs nm ? |
26 |
>>>>>> How far is cache size important ( 6 vs 8 MB )? |
27 |
>>>>>> |
28 |
>>>>>> When I built my current machine 2007, the CPU cost CAD 213 , |
29 |
>>>>>> so both look as if they're in the right ballpark. |
30 |
>>>>>> |
31 |
>>>>> If you're building new, performance-oriented box, you should take latest |
32 |
>>>>> intel with AVX because of AVX. As I recall, recent gcc has support for |
33 |
>>>>> avx, |
34 |
>>>>> so some performance gain may be achieved. |
35 |
>>>>> If you want home box, you may be interested in AMD A8 and similar chips, |
36 |
>>>>> as |
37 |
>>>>> they are reasonably fast and very chip |
38 |
>>>> |
39 |
>>>> AMD parts have had AVX since the Bulldozer core release in Q3 2011. |
40 |
>>> |
41 |
>>> Are they already available in reasonable numbers on market? |
42 |
>> |
43 |
>> http://www.cpubenchmark.net/cpu.php?cpu=AMD+FX-8120+Eight-Core |
44 |
>> |
45 |
>> At $150, fitting into existing Socket AM3+ boards, that looks like the |
46 |
>> best part for my money right now. |
47 |
>> |
48 |
>>>>> In any case, I'd put most of my money in 2-4 big 3Tb HDD's for media and |
49 |
>>>>> 8+ |
50 |
>>>>> Gb fast memory, as modern browsers eat memory like crazies and CPU is |
51 |
>>>>> usually fast enough. Decoding HDTV mkv's should occur on gpu block in any |
52 |
>>>>> case, so general performance for most uses is irrelevant, as it was fast |
53 |
>>>>> enough four yesrs earlier. Simply check, that you can offload HDTV |
54 |
>>>>> decoding |
55 |
>>>>> to GPU in your config. |
56 |
>>>> |
57 |
>>>> Here, you're talking about either VDPAU or VAAAPI support. VDPAU is |
58 |
>>>> only offered by nVidia cards, and even then you need to run the |
59 |
>>>> proprietary driver. VAAPI is supported by Intel graphics and ATI's |
60 |
>>>> proprietary driver. |
61 |
>>> |
62 |
>>> I do not see any problems with this. A blob in system is not best practice, |
63 |
>>> of course, but it does not need any configuration and is not a performance |
64 |
>>> bottle-neck, so there is no reason to care. |
65 |
>> |
66 |
>> I only bring it up because some people do care. I'm running fglrx at |
67 |
>> home right now. When I run nVdia, I run the nVidia drivers. In part |
68 |
>> because I like accelerated video decoding (which a Geforce 210 does |
69 |
>> wonderfully), in part because the nv, nouveau and radeon drivers |
70 |
>> historically worked very poorly for me in 2D performance when faced |
71 |
>> with multiple 1080p displays. They're always getting better, of |
72 |
>> course. |
73 |
>> |
74 |
>>> |
75 |
>>> I personally would prefer AMD A8 if I can offload decoding to GPU unit there |
76 |
>>> (not sure if I can, so won't change my box till next summer), but discrete |
77 |
>>> video card will not be the most costly part in good non-gaming box, hard |
78 |
>>> drives will, so again, what the matter? |
79 |
>> |
80 |
>> Computer usage breaks down into more than gaming and non-gaming. My |
81 |
>> "non-gaming" boxes at home tend to have their CPU, RAM or NICs as |
82 |
>> their most expensive components, because that's where I need them to |
83 |
>> perform better. |
84 |
>> |
85 |
> |
86 |
> CPU speed does not matter. what matters most is the I/O speed. |
87 |
> |
88 |
> As far as I can tell, AMD chip suffered with a lot of I/O. Their |
89 |
> Hyper-transport seems not competitive with Intel's ring bus |
90 |
|
91 |
(please don't top-post, especially if the thread's already been |
92 |
primarily organized as bottom-post) |
93 |
|
94 |
I hadn't read that, but remember that HyperTransport is intended for a |
95 |
mesh architecture. In single-CPU systems, you'll only have one HT |
96 |
link, the link between your CPU and your north bridge. In multi-CPU |
97 |
systems, you'll have additional links between the CPUs. In systems |
98 |
with many CPUs, you may even have a fully-connected mesh. |
99 |
|
100 |
The I/O characteristics will greatly depend on the topology of your network. |
101 |
|
102 |
That said, HyperTransport may just be getting old; when it came out, |
103 |
it (and AMD's crossbar switch for memory management) beat the pants |
104 |
off of Intel's SMP solution. Intel's solution ran at lower and lower |
105 |
clock rates the more CPUs you added, and their first pass at multicore |
106 |
gave each core its own port onto the memory bus, with predictably poor |
107 |
results. Intel's had plenty of time to catch up, but with their |
108 |
price-per-part, it's taken me a long time to pay much attention. |
109 |
|
110 |
(It also doesn't help that Jon "Hannibal" Stokes stopped writing |
111 |
detailed technical articles for Ars Technica; I sincerely miss him and |
112 |
the precision and clarity of his writing on such arcane subjects.) |
113 |
|
114 |
-- |
115 |
:wq |