What I have Learned about Computer Performance from SPCR

The forum for non-component-related silent pc discussions.

Moderators: NeilBlanchard, Ralf Hutter, sthayashi, Lawrence Lee

MikeC
Site Admin
Posts: 12285
Joined: Sun Aug 11, 2002 3:26 pm
Location: Vancouver, BC, Canada
Contact:

Re: What I have Learned about Computer Performance from SPCR

Post by MikeC » Wed Feb 23, 2011 2:13 pm

ces wrote:
MikeC wrote:Snobbery against integrated video is a long tradition among performance-oriented users, but not nearly as justified as before. If you are not a gamer, current integrated video from either AMD or Intel work perfectly well for just about everything. Sandy Bridge HD2000 is now the best performance integrated graphics available, and identical to HD3000 except for 3D (gaming), so you can save your $ if that's the only reason you consider the 2500K.
This is very interesting info. Can anyone tell me how to look at a review and understand what kinds of non-gaming performance an integrated or non-integrated video chip set will do. I have not seen info in reviews that would let me in on what aspects of performance are non-gaming related. I had thought in fact that the Sandy Bridge HD3000 was all around faster than the Sandy Bridge HD2000... not just in gaming.

Can anyone help me understand this?

For instance, can you help me read this review, comparing the Sandy Bridge HD3000 to the Sandy Bridge HD2000, to better understand where they are different and where they are not?
http://www.xbitlabs.com/articles/video/ ... -3000.html
Intel HD Graphics 3000 has twelve "execution units", while the 2000 has has six. This is only relevant with 3D effects. The 2D features are identical. The 3000 is also often clocked faster (depending on CPU model), and it is mated to faster/unlocked CPUs. That xbit labs review looks mostly at 3D performance, where the HD 3000 (with its higher clock and faster mated CPU) always has a big edge over the HD 2000. But look at the general (non-3D) performance benchmarks (like PCMARK) on page 6 of the xbitlabs review, and you will see that here, the differences between 3000/2000 are trivial. But it's hard to do an absolute a/b with 3000/2000 on either 2D apps (like photoshop) or 3D because the CPU performance matters, and HD 2000 is usually mated to a slower CPU than HD 3000.

tim851
Posts: 543
Joined: Wed Aug 13, 2008 11:45 am
Location: 128.0.0.1

Re: What I have Learned about Computer Performance from SPCR

Post by tim851 » Wed Feb 23, 2011 3:39 pm

ces wrote:I had thought in fact that the Sandy Bridge HD3000 was all around faster than the Sandy Bridge HD2000... not just in gaming.
There is not really anything much beside gaming. That's why it is so easy to recommend integrated graphics to people who don't game.

The HD3000 will be faster than the HD2000 in anything 3D-related, not just games. They're both outlandishly slow though, so whenever you need to do something 3D, you will want a discrete card.

Then there's also GPGPU, i.e. software that uses the GPU rather than the CPU for general computing tasks. Nvidia has this field covered with CUDA. nVidia and AMD both implement Apple's OpenCL, which should also work on Intel GPUs. So if you have software that uses the GPU for certain things, the HD3000 should also be a little faster.

Everything 2D related, including video decoding, will perform identically on the two iGPUs.

ces
Posts: 3395
Joined: Thu Feb 04, 2010 6:06 pm
Location: US

Re: What I have Learned about Computer Performance from SPCR

Post by ces » Wed Feb 23, 2011 4:35 pm

MikeC wrote:Intel HD Graphics 3000 has twelve "execution units", while the 2000 has has six. This is only relevant with 3D effects. The 2D features are identical. The 3000 is also often clocked faster (depending on CPU model), and it is mated to faster/unlocked CPUs. That xbit labs review looks mostly at 3D performance, where the HD 3000 (with its higher clock and faster mated CPU) always has a big edge over the HD 2000. But look at the general (non-3D) performance benchmarks (like PCMARK) on page 6 of the xbitlabs review, and you will see that here, the differences between 3000/2000 are trivial. But it's hard to do an absolute a/b with 3000/2000 on either 2D apps (like photoshop) or 3D because the CPU performance matters, and HD 2000 is usually mated to a slower CPU than HD 3000.
So for purposes of non-3d/non-gaming video performance, the place to look is the PCMark test results like these:
http://www.xbitlabs.com/articles/video/ ... html#sect0

I went to look at Tom's hardware's graphics testing. They don't appear to use PCMark. Are any of their tests useful for evaluating non-3d/non-gaming video performance?
http://www.tomshardware.com/charts/2010 ... s,114.html

tim851
Posts: 543
Joined: Wed Aug 13, 2008 11:45 am
Location: 128.0.0.1

Re: What I have Learned about Computer Performance from SPCR

Post by tim851 » Wed Feb 23, 2011 5:55 pm

ces wrote:So for purposes of non-3d/non-gaming video performance, the place to look is the PCMark test results like these:
http://www.xbitlabs.com/articles/video/ ... html#sect0
Well, they test the CPU. I don't think the GPU plays a role there at all.
I went to look at Tom's hardware's graphics testing. They don't appear to use PCMark. Are any of their tests useful for evaluating non-3d/non-gaming video performance?
http://www.tomshardware.com/charts/2010 ... s,114.html
No. Except for 3Dmark, all their benchmarks are actual game benchmarks. And 3Dmark is just a synthetic video-game-benchmark.

"Video performance" can be quantified by corresponding CPU utilization, as done in most SPCR review. Some video-centric websites might also test what kind of post-processing effects you can apply. That's one of the reason to go with a Radeon 5550 over the 5450, as far as I know. The latter supports more/better post-processing.

ces
Posts: 3395
Joined: Thu Feb 04, 2010 6:06 pm
Location: US

Re: What I have Learned about Computer Performance from SPCR

Post by ces » Thu Feb 24, 2011 3:20 pm

Fan Cooling Recipe:
MikeC wrote: A. Objective: To achieve good cooling of core components with the absolute minimum audible noise in a PC. This ignores cost.
B. Assumption: At least 1 fan in the system & the ability to control the speed of each fan. Heatsinks are close to ideal for low airflow cooling performance.
C: Fan options: All 120mm fans, potentially up to 8. (Just to set a guideline) 1 on pack panel, 1 on top panel (like in P183), 2 on front panel, 2 on CPU HS, 2 on VGA HS.

1. Start with the minimum number of fans you believe is needed. Choose fans on the basis of best noise/rpm ratio at <16 dBA/1m by SPCR testing.
2. Tweak the speed of each fan by ear/temp while applying stress utility appropriate to your intended usage. If satisfactory noise/temps are reached, congratulate yourself, have a drink.
3. If not satisfied, add a fan somewhere -- or perhaps 2 -- and redo the process. Focus on getting the overall noise of the fans as quiet or quieter than before, while improving or maintaining temps (depending on whether temps or noise was the issue before). If satisfactory noise/temps are reached, congratulate yourself, have two drinks.
4. If not satisfied, repeat as necessary... or re-examine your fan selection and consider changing to another model of fan. Have drink to kill the pain.

ces
Posts: 3395
Joined: Thu Feb 04, 2010 6:06 pm
Location: US

Re: What I have Learned about Computer Performance from SPCR

Post by ces » Tue Mar 01, 2011 6:31 am

leem wrote:In terms of idle power consumption, is Sandybridge really that much better?

The Intel DH67BL is 17W at idle, but so is the Intel DH57JG mini-ITX board.
I must say that is very interesting. Though the sandy bridge on the DH67BL does get the same idle wattage with 4 cores as the Clarkdale gets with 2 cores on the DH57JG. Still... that is darn close all the same.

ces
Posts: 3395
Joined: Thu Feb 04, 2010 6:06 pm
Location: US

Re: What I have Learned about Computer Performance from SPCR

Post by ces » Thu Mar 03, 2011 12:13 pm

CPU Coolers for Real Men

Scythe Susanoo CPU Cooler Has Four 120mm Fans And Twelve Heatpipes

Read more: http://www.madshrimps.be/news/item/77658#ixzz1FZKRZI9K

Notice the similarity to the Parrot Drone
http://ardrone.parrot.com/album

Oh and don't forget
http://www.nofencomputer.com/eng/


Last edited by ces on Tue Apr 24, 2012 1:42 pm, edited 4 times in total.

MikeC
Site Admin
Posts: 12285
Joined: Sun Aug 11, 2002 3:26 pm
Location: Vancouver, BC, Canada
Contact:

Re: What I have Learned about Computer Performance from SPCR

Post by MikeC » Thu Mar 03, 2011 12:23 pm

ces wrote:CPU Coolers for Real Men
Not really. It's about the wow factor, getting news coverage at trade shows. Scythe has a looooong history of doing this over the years -- ridiculousy big HS introduced at a show, lots of headlines, but few sales and sometimes, no retail product. Not so different from Thermalright lat times -- last one I recall are those heatsink cases from a couple years back.

Everybody's doing it tho. Trying to get free PR, wow the techheads.

ces
Posts: 3395
Joined: Thu Feb 04, 2010 6:06 pm
Location: US

Re: What I have Learned about Computer Performance from SPCR

Post by ces » Mon Mar 07, 2011 11:31 am

deleted by author
Last edited by ces on Sun Mar 27, 2011 6:37 am, edited 4 times in total.

ces
Posts: 3395
Joined: Thu Feb 04, 2010 6:06 pm
Location: US

Re: What I have Learned about Computer Performance from SPCR

Post by ces » Mon Mar 07, 2011 11:49 am

chanman wrote:JonnyGuru recently reviewed a few pico PSU/brick combinations
http://www.jonnyguru.com/modules.php?na ... y&reid=207

ces
Posts: 3395
Joined: Thu Feb 04, 2010 6:06 pm
Location: US

Re: What I have Learned about Computer Performance from SPCR

Post by ces » Wed Mar 09, 2011 9:27 pm

issues on Asus P67 boards
cyreb7 wrote:
ces wrote:What problems did you find people having?
Lots of people had trouble with SSDs and some had weird boot issues on Asus P67 boards, there was other stuf as well but I don't remeber. You might want to skim through this thred.
Last edited by ces on Wed Mar 23, 2011 3:03 pm, edited 2 times in total.

ces
Posts: 3395
Joined: Thu Feb 04, 2010 6:06 pm
Location: US

Re: What I have Learned about Computer Performance from SPCR

Post by ces » Wed Mar 09, 2011 10:56 pm

Differences between Intel G2 SSDs
quest_for_silence wrote:
Jenny wrote:Intel SSDSA2MH080G2R5 ($245)
OR
Intel SSDSA2MH080G2K5 ($261)
As far as I know the differences should be the following: if the product code ends in "Kn" (where "n" is a number), then it's a "Retail kit" (which usually comes with SATA cable, the 3.5 to 2.5 adapters, a 4p Molex to SATA-power adaptor, and Intel "bragging rights" stickers). On the contrary, if the product code ends in "Rn" (where n"" is a number), then it's a "Retail standard box" (which usually comes with a slightly lighter bundle, I mean just the 3.5 to 2.5 adapters, and the same Intel official stickers).
Last edited by ces on Wed Mar 23, 2011 3:02 pm, edited 2 times in total.

ces
Posts: 3395
Joined: Thu Feb 04, 2010 6:06 pm
Location: US

Re: What I have Learned about Computer Performance from SPCR

Post by ces » Sat Mar 12, 2011 2:17 pm

Quiet Keyboards
MikeC wrote:Obviously, if you haven't, you should take the time to scroll through the massive 439-post, 9-year thread Recommend a Silent Mouse and Keyboard? -- viewtopic.php?f=23&t=2500
For info on quiet Keyboards see
viewtopic.php?f=23&t=2500
and
viewtopic.php?f=1&t=61769&p=536148
and
viewtopic.php?f=1&t=61778&p=536166

andymcca wrote:
ces wrote:Neat tiny keyboard for HTPC
Cideko AVK-02-915 Black RF Wireless Mini Keyboard
http://www.newegg.com/Product/Product.a ... 6823721001
http://www.newegg.com/Product/Product.a ... 6823721001
(link was broken!)
Ask Engadget: best sub-$150 mechanical keyboard?
http://www.engadget.com/2012/04/21/ask- ... -keyboard/
Last edited by ces on Sun Apr 22, 2012 7:46 am, edited 3 times in total.

ces
Posts: 3395
Joined: Thu Feb 04, 2010 6:06 pm
Location: US

Re: What I have Learned about Computer Performance from SPCR

Post by ces » Mon Mar 14, 2011 7:02 am

Quiet Hard Drives
fumino wrote:if you only have that one hard drive, then look into the "NoiseMagic NoVibes III" (old spcr link, theyre still being made)
with that, you just have to find a decent midtower. just gotta find a balance of looks, performance, features and price... and how much it will actually take/cost to make it quiet.
NoiseMagic NoVibes is a perfect solution for anyone who wants suspended hard drives but for one reason or another doesn't want to use the Antec Solo case.
MikeC wrote:
rpsgc wrote:Need opinions on the Sharkoon Vibe Fixer. Considering it looks exactly like a NoVibes, can one assume it's as good?
yes
Last edited by ces on Sun Jun 12, 2011 10:36 am, edited 1 time in total.

ces
Posts: 3395
Joined: Thu Feb 04, 2010 6:06 pm
Location: US

Re: What I have Learned about Computer Performance from SPCR

Post by ces » Mon Mar 14, 2011 10:13 am

Top CPU Coolers
ces wrote:
dhanson865 wrote:Here is a quote of a post I made in another thread
Given the new review I thought I'd do a quick price check of the top half of the test results on the temp rise chart. I dumped the ones that were overpriced vs the performance or had mounting issues and came up with this short list. Prices were from a quick google search and include S&H and taxes for my zip code in the US.

Thermalright Ultra-120 eXtreme Rev.C $66
38 degree rise 790g
Mounts both ways on AMD boards. Editor's Choice, best on SPCR list at end of 2010. Gets a 9.5 on the recommended list.
Noctua NH-U12P $64
39 degree rise 600g without fan / 770g with stock fan
Mounts both ways on AMD boards. Good choice for those that move their PC a lot and are afraid of heavier heatsinks. Gets a 9 on the recommended list.
Scythe Mugen-2 $44
39 degree rise 870g including fan
Mounts with "proper" orientation on AMD boards but only mounts in that one direction. Gets a 9 on the recommended list.
Cooler Master Hyper 212 Plus $29 (best bang for the buck)
41 degree rise 626g including fan
Mounts both ways on AMD boards

Again the Cooler Master 212 Plus is light enough for those afraid of heavier heatsinks. Gets a 8.5 on the recommended list because the stock fan throbs but for a SPCR regular this is easy to fix with a fan swap. Grab your favorite 12cm/120mm fan and make it right.


That list was made before the Noctua NH-C14 review and the NH-C14 hasn't been added to the recommended list. Noctua NH-C14 is $90 shipped to my house so it's certainly not a value heatsink but it is the best top down heat sink and it's very flexible for low clearance and high clearance situations.

Best value only makes sense if you do a search based on the price it takes to get it in your hands and that includes retail price + shipping + tax/VAT + any odd fee someone might tack on + real world considerations like driving to a post office or store or whatever you personally might have to get that device in your hand.

It also matters what hardware you have laying around or already in your PC. If you have a stack of high quality fans you might consider the Hyper 212 Plus a good value, if you have crappy fans and need to step up the Mugen-2 might be cheaper for you than buying a Hyper 212 Plus and buying a better fan.

Only you can decide how much a heat sink would cost you given all the possible variables.
If you are going to buy a low TDP CPU you could use any decent HS fanless. I'm using a X3 720 89W TDP cpu with a Gelid Tranquillo and I just turned the CPU fan off a few minutes ago while typing this. CPU temp hasn't rise much from the mid 30s it normally idles at with the fan running around 350RPM. I could easily leave the CPU fan off if I were using a lower TDP chip or if I were willing to undervolt/underclock a little.

If you had the Thermalright Ultra-120 eXtreme Rev.C it'd work even better than this Tranquillo.

If you want the one with the best fin spacing that's the Noctua NH-U12P. Basically the same fin spacing as a Ninja 3 but much better cooling

Code: Select all

Heatsink                             Fin Spacing
[b]Noctua NH-U12P[/b]                       2.63 mm
[b]Scythe Mugen-2 [/b]                      1.89 mm
[b]Cooler Master Hyper 212 Plus[/b]         1.54 mm
[b]Thermalright Ultra-120 eXtreme Rev.C[/b] 1.52 mm

[b]Noctua NH-C14[/b]                        1.79 mm (horizontal)
dunno how the top down would play out in fanless mode but depending on the orientation in your case it could be a plus for passive cooling.
Noctua Ram Clearance
Rasta_Cook wrote:b]CPU COOLER / HEATSINK:[/b]
1.Noctua NH-U12P SE2 (link):
This heatsink use a tower design. I was thinking I could use this cooler with only 1 fan, and use the other fan somewhere in my case. The noctua fan would probably be quieter than the stock fans coming with whatever case i choose, and with 1 fan i think the heatsink is good enough to keep the slightly overclocked cpu within acceptable temps. Problem with this cooler is that the clearance for the ram is only 40mm and i dont know if that is enough for the g.skill ripjaw x. The noctua fans come with 2 adapters to control speed. I plan on using the medium speed if possible ( LNA adaptor ). height of this cooler is 158mm

2. NH-C12P SE14 link:
This heatsink uses a top flow design. It sends air on the motherboard component which is good i guess. It doesnt come with 2 fans though but it is cheaper than the NH-U12P therefore I can purchase another noctua 120mm fan for my case if i want. This heatsink uses a 140mm fan instead of a 120mm fan, Therefore should be more quiet due to fan being bigger. The main problem with this heatsink is that it would disrupt the airflow in the case, im not sure if this is game-breaking or not, but I think the NH-U12P would be more effective at keeping warm air out of the case since it sends hot air directly at the case rear exhaust fan. Ram clearance is 44mm. so only 4mm more than the NH-U12P.

3. NH-C14 link :
so for maximum ram clearance this heatsink offers 65mm. It also comes with 2 140mm fans, but i will only use 1, the other one will be used in the case if the case has a 140mm fan to replace. This is only in my consideration only because of potential ram clearance issues. I think the NH-C12P is a better single fan top flow because the heatsink fins extend all the way to the bottom of the cpu heatspreader.
Last edited by ces on Sun Mar 27, 2011 6:38 am, edited 1 time in total.

ces
Posts: 3395
Joined: Thu Feb 04, 2010 6:06 pm
Location: US

Re: What I have Learned about Computer Performance from SPCR

Post by ces » Mon Mar 14, 2011 2:26 pm

XP to Windows 7 Upgrade
DaveLessnau wrote:The Microsoft page for a clean install of Windows 7 over an existing XP setup is here:

http://windows.microsoft.com/en-us/wind ... -windows-7

The meat of the thing is under the Step 3 tab on that page. An alternate rendition of those instruction would be:

http://www.winsupersite.com/article/win ... media.aspx

That version even talks a bit about the Family Pack and the double-install you mentioned.

Don't forget to do full system images and backups if at all possible before doing this. That way you can recover if things go bad. Also, to avoid all of this dance if you ever have to wipe your installation and start over, get a full system image done on your system as soon as it's set up and store that image someplace safe.

As far as OS reactivation prompts, I can't answer (since I'm not running an OEM version). In my case, I've only seen one once when I changed motherboards. Simply running the Activate Windows tool fixed it. I also had Office 2007 (but Windows 7 didn't care about it) hit me with a re-activation request when I added memory. Again, no problem.

Also, you might have to do some finagling if you're moving from x86 to x64 versions of OSes. Read here:

http://www.winsupersite.com/article/win ... ows-7.aspx

And, as far as pricing, Amazon is selling what *appears* to be the retail (i.e., non-OEM) upgrade version for $106:

http://www.amazon.com/Microsoft-Windows ... 025&sr=1-2

I can't guarantee that it's retail instead of OEM, though.
andyb wrote:The "retail" AKA "I am a mug" licence is off-limits due to the stupid cost with no benefit.

Upgrades are never a great idea if you actually have to install another OS then actually upgrade. I did hear some rumours some time ago about being able to do a "clean-install upgrade" but I never found out if it became reality.

The "family-pack" is certainly cheaper if you need 3 licences, the obvious points of interest are that you really want the 64-bit version, that means that all 3 machines need to be 64-bit capable, but additionally it should make it easier (in theory) to install 2 machines now, then when you come to "upgrade" one of those machines you can use the 3rd licence to do it, then use one of the original licences again later.

One of the main problems with the shitty licencing bullcrap that MS force upon the people who give them cash is that if you install the OS, then activate it, then you find out that your brand new mobo/HDD is faulty and it trashes the OS they might not like you re-activating the same code again in a short time space - I have run into this problem before and ended up having to put a non-legal OS on the customers machine (with their blessing) because of this - so be wary.

As far as the actual triggers of re-activation, its pretty weird and random. Although as a general rule your motherboard (almost) always counts, and if your machine is made by a big OEM your HDD can also count, CPU, RAM, PSU, Keyboard, Mouse and other external devices no, everything else is random but rare, so basically its the motherboard in most home builds that counts, but even then if its say 6-months after you first activated your copy of W7, and you install the same licence on a new PC it might activate staright away with no questions even though it is still activated and working on the old PC - this is of course dangerous if you have Automatic Updates turned on, and you might find that your licence gets red-flagged and deactivated.

There is one further option if you are concerned, get your licence and dont activate it, use a naughty method until you replace your PC then simply re-install using your legitimate code take an image of the drive once you are happy and everything will work just fine.

Finally, there is a little trick when installing W7, choose to create a partition, it will then tell you that it will create a 100MB partition for some reason that I dont care about, and then will make the "main" partition after that. What you do is then delete the "main" partition, and extend the 100MB partition to your desired size, simply finish the install as you would do otherwise.


Andy
SebRad wrote:Hi, the copies of Windows 7 FP that I have have both 32 and 64bit disks in them.
It IS an upgrade licence BUT easy enough to get round. you mentioned you have XP installs so just install XP minimally, like to the first full boot, don't worry about drivers or anything else at all, on the machine (or leave the original install, especially if working)

Boot off the Windows 7 DVD and during the "Upgrade" process you can tell it you want to re-partition/format the HDD so your old Windows is blow away anyway.
You can't directly upgrade a x86 OS to x64 anyway. I've done several like this with no problem at all. In fact one of them was an XP install from another machine, just transferred the HDD (which wouldn't then even boot due to HW changes) put in the upgrade DVD and away it went!

I also managed to "upgrade" (with bit registry hacking) a dodgy copy of Windows 7 Ultimate x64 with W7 HP FP x64 and it wouldn't activate itself. I had to call up the phone number and ended up speaking to someone (nice man in India I think!) and confirmed was Family pack and no this was only the 3rd use of it and was given activation code and all been working fine for months since.
XP and Windows 7 Compatibility
Plekto wrote:... be aware that the Pro version has XP compatibility more while the lesser versions are crippled and don't have it.
SebRad wrote:Hi, almost all applications (especially MS ones) will install just fine on all versions of Windows 7. They all have "compatability" settings to try to fool the program it's on an older version of Windows for older / troublesome apps. I think in some case 32 bit will have less trouble than 64bit.

The "XP Mode" is aimed at legacy or cutome business applications that are [badly] programed in such a way as they will only work under Windows XP, and will not work with the "normal" compatability settings.
What the XP mode gives you (I think) is a well integrated virtual machine. Effectively a virtual machine is run and runs the application and is then presented in a normal window.

You can achieve a similar effect with MS Virtual PC, (thats free - and is what's used by XP Mode!) if you have a copy of XP you can install on it.

I've have Windows 7 x64 and had no problems with pretty much anything installing on to it.

For other people the only thing I've had to resort to a virtual machine for was an old Sage package that would only almost run under Windows. Just created a virtual PC with minimal XP install and Sage in the startup group and created a shortcut to the VM on the desktop so it's pretty much one click to open it
DaveLessnau wrote:What he means by that is that Pro, Ultimate and Enterprise come with a virtual machine for XP. IOTW, you can fire up a virtual PC of XP from within Windows 7 and run any particular piece of software that won't run under Windows 7 but did under XP (I've never run into anything like that). Here's the link:

http://windows.microsoft.com/en-US/wind ... ws-xp-mode
ces wrote:Does that mean that my MS Word 2007 won't install on my Home version of Win 7?
DaveLessnau wrote:No. Word 2007 will install just fine under Windows 7.
Last edited by ces on Fri Mar 18, 2011 8:09 am, edited 1 time in total.

ces
Posts: 3395
Joined: Thu Feb 04, 2010 6:06 pm
Location: US

Re: What I have Learned about Computer Performance from SPCR

Post by ces » Tue Mar 15, 2011 9:08 am

Deleted by Author
Last edited by ces on Wed Mar 30, 2011 8:42 pm, edited 3 times in total.

ces
Posts: 3395
Joined: Thu Feb 04, 2010 6:06 pm
Location: US

Re: What I have Learned about Computer Performance from SPCR

Post by ces » Tue Mar 15, 2011 8:51 pm

The Stack Effect With Pringles Can
mb2 wrote:i3-530 @ 2GHz, prime95 for 75 mins.
Hiper 212-plus. No fans: = 66/7 C
With the help of a pringles can, some cardboard and a little duct tape: = 56/7 C
10C fall not so bad for a couple of minutes bodge job eh?

ces
Posts: 3395
Joined: Thu Feb 04, 2010 6:06 pm
Location: US

Re: What I have Learned about Computer Performance from SPCR

Post by ces » Mon Mar 21, 2011 4:00 pm

The "How To" of Case Modifications

Modding & Cutting with Tin Snips
viewtopic.php?f=15&t=2307

The basics
viewtopic.php?p=3063#p3063\

How to buy snips
viewtopic.php?p=17478#p17478
viewtopic.php?p=17692#p17692
viewtopic.php?p=427873#p427873

Basics on Tools
viewtopic.php?p=36397#p36397

Cutting Holes

If you want to cut a hole for a fan, go to ebay and purchase the hand "brace" drill. Using a power drill with the crude hole saws that are available just doesn't work well. The teeth kick back and you are at risk of wrenching or even breaking your... it it just plain chews up the metal. Go visibly inspect the shape and direction of the teeth on them and you will see what I mean.
80mm fan = 3.14960 inches = 3 inch hole saw
92mm fan = 3.62204 inches = 3.5 inch hole saw
120mm fan = 4.72440 inches = 4.5 inch hole saw

Here is how the pros do it:
Knockout punches to punch a 120mm hole
Greenlee Knock out Punches
Greenlee part # 730EBB-120 Estimated purchase price of $550 new.
http://www.greenlee.com
www1.mscdirect.com
themaster1 wrote:I have asked the same question to a guy who made a neat mod (cut a hole on aluminum) a while ago and he advised me to use a this tool: http://www.youtube.com/watch?v=9iMiH8wYTDY (cutting hole demo at 2.46) You cut a small hole first and then at the right diameter it's easier. The sharp edges can be dealt with sandpaper (double sided)
The Effect of Grills on Noise and Performance
lapon wrote:Check out this link about how grills affect fan noise and performance. They were testing the grills right next to the fan, but it might help to pick a good one since the case side is close anyways. Alternatively, can you just run with the case open? If the CPU/PSU/GPU fans are relatively quiet its not a bad temporary fix and you don't need to have the case fans on anymore since the case is wide open.

The Best Hole Cutter for Non-Pros

One device that works reasonably well at a fairly low cost is the Malco HC1 - 2" to 12" Sheet Metal Hole Cutter Accessory
http://www.comfortgurus.com/product_inf ... ts_id/2000
http://malcoproducts.com/product/hvacr/ ... le-cutters
Last edited by ces on Mon Jul 16, 2012 8:42 am, edited 7 times in total.

ces
Posts: 3395
Joined: Thu Feb 04, 2010 6:06 pm
Location: US

Re: What I have Learned about Computer Performance from SPCR

Post by ces » Tue Mar 22, 2011 5:44 am

Mounting Fans with Those Rubber Thingees - Which One Are the Best?
flemeister wrote:
jcoffin1981 wrote:Does anyone have any experience with using the rubber grommets to mount case fans? They are a freaking pain in the arse to put in and after ripping a few just screwed it instead. The Noctua fans hardly seem to vibrate. There may not be an advantage to this technique. What are your thoughts? Thanks!
You're using crap ones. Try the ones that Nexus include with their Real Silent fans, they're the gooiest, softest and toughest ones I've come across so far. :)

ces
Posts: 3395
Joined: Thu Feb 04, 2010 6:06 pm
Location: US

Overclocking the i7-2600K

Post by ces » Wed Mar 23, 2011 3:00 pm

Overclocking Sandy Bridge
Mats wrote:When Anandtech overclocked the 2600K from 3.4 GHz to 4.4 GHz, the power consumption raised 25 W.
Overclocking the Sandy Bridge, doesn't burn up that much energy (or generate that much extra heat) It doesn't affect idle heat or temps. It only comes into play when you are pushing the top end. That is very hard to do with a 2600K for any sustained period. About the only way you can do it is with benchmarking programs.

Most of the time it will be operating at idle or near idle. That is where most of your energy will be burned... and it will be burned at the same rate, generating the same heat, as a non-overclocked CPU.

ces
Posts: 3395
Joined: Thu Feb 04, 2010 6:06 pm
Location: US

Sandy Bridge Graphics HD2000 vs HD3000

Post by ces » Fri Mar 25, 2011 7:28 pm

Sandy Bridge Graphics HD2000 vs HD3000
boost wrote:The maximum resolution is 2560x1600 according to this slide.
MikeC wrote:In case you did not notice, there is no difference in non-3D performance between the HD2000 and HD3000. Intel states that explicitly, and we confirmed it. The differences show up only in 3D apps like games.
Vicotnik wrote:
lurpitus wrote:Let's assume I'll get 30" screen with that resolution. Is the HD 3000 really enough for it for non gaming use? I would assume that playing any kind of HD content full screen would require more power than what HD 3000 offers. I'm thinking about either of the following setups:
1x30" + 1x24"
2x24"
Your HD content have a fixed resolution. Decoding is the hard part, scaling the image to your resolution is not a very complicated process and 3D performance has little to do with it. So you'll be fine. :)
gb115b wrote:missingremote is the only one that talks about 24fps support with UAC disabled, you'd probably have to ping them for more details.

Anandtech claims in their article that their is no 24p support due to a limitation in the chipset (from what i read)..

I guess only Intel know the real answer! judging by missing remote it looks liek something that could come to a future driver? but what they actually had was not 23.976, but something close 23.974 (I think?) which will also judder...though a lot less?
Intel’s Sandy Bridge Celeron class CPUs are missing some video features
(a) no hardware assisted video encode
(b) no post encoding processing such as adaptive contrast, skin tone, color control
http://semiaccurate.com/2011/03/21/inte ... fused-off/
Last edited by ces on Sat Mar 26, 2011 6:12 am, edited 1 time in total.

ces
Posts: 3395
Joined: Thu Feb 04, 2010 6:06 pm
Location: US

Differences between the older X-x50 and newer X-x60 Seasonic

Post by ces » Fri Mar 25, 2011 7:33 pm

The difference between the older X-x50 and newer X-x60 Seasonic PSUs
dhanson865 wrote:* Larger/more/better heatsinks
* Better voltage regulation, ripple, and such
* cleaner PCBs redesigned with the benefit of time vs the first gen X-series

maybe other pluses but afaik it's all improvements on the X-x60 vs the old X-x50 units.

For pictures of the internals see page 4 of the JonnyGuru reviews.

http://www.jonnyguru.com/modules.php?na ... 4&reid=169

http://www.jonnyguru.com/modules.php?na ... 4&reid=202

oh and "ferky" on the forums at JG listed these differences between the 560 and 650. Go to the seasonic website and compare the 560 to the 660 to see how they compare if the 560 to 650 comparison confuses you.

* X-560 has few cm longer cables than X-650, 2x PCI-E 6+2 (X-650 has 4), 1x 3,5" floppy (X-650 has 2), 1x 8pin EPS12V (X-650 has 2 + 1x 4pin). 5x 5,25" Drive connector (X-650 has 8 )
* X-560 has a bit stronger fan (difference is 200-300RPM @12V), but probably not louder (because the PWM kicks in)
* Transient stage is a bit rearranged (positions of MOV, thermistor and fuse are "hidden" behind the relay)
* Heatsinks on the primary stage are colored in orange/gold on X-560
* PFC Mosfets on the X-560 are assembled in a TO-220 case which has an substantial advantage over TO-220FP (FullPak) in X-650 (in aspect of power dissipation).
* Mosfet trasistors on the switching section of the X-560 have a bit lower typical RDS(on) then the ones inside X-650 (which has a bit powerfull ones).
* PWM controller (CM6901) on the X-560 is no longer placed on the main PCB but on a small daugtherboard (opposite of the X-650).
* The +3.3/5V secondary Mosfet transistors in X-560 have a lower RDS(on) and have higher current limits on high temperatures (and also higher peak/pulse values).
* Some of the electrolytic capacitors on the secondary have higher values in X-560 (than the ones in X-650).
* Secondary +12V Mosfets in X-560 have better heatsinks (lower °C/W) than the ones in X-650.
* The PS223 monitoring integrated circuit in the X-560 has a bit of a help from a couple of LM393 Voltage Comparators (which isn't the case on X-650).
* Some smaller transformers are different between those two PSU (nothing too important).

folowed by the comment "all those improvements are really not important (except maybe for number of cables/connectors in X-650, if they are needed)."

Short answer is buy whichever is cheaper unless you want to buy the latest and greatest in which case get a x60 wattage.

ces
Posts: 3395
Joined: Thu Feb 04, 2010 6:06 pm
Location: US

Re: What I have Learned about Computer Performance from SPCR

Post by ces » Mon Mar 28, 2011 7:50 pm

Fan Tests
mdrumt wrote:Don't know if this has been posted anywhere before.
Has anybody else seen this fan review site? Looks pretty new.
http://atreview.net/en/casefan.html
Each fan has a Youtube video "review" with noise / cfm measurements and quality audio recordings!
ces wrote:Here is an interesting post:

Fan curves: I can make them
viewtopic.php?f=9&t=61225
ces wrote:How about testing this fan.

Scythe Gentle Typhoon 120mm x 25mm Fan - 3000 RPM (D1225C12B7AP-29)
83.0 CFM at 36.5dBA

http://www.frozencpu.com/products/12363 ... _Soon.html

Not exactly quiet, but interesting nonetheless.
ces wrote:Here is the specs on the family of NMB fans that Dorothy Bradbury sells:
http://www.nmbtc.com/cgi-bin/start.cgi/ ... _size=4.69
These are the two 120mm models she sells:
4710KL-04W-B10
4710KL-04W-B50
(her exact models may have slightly higher grade bearings and electronics)

If you click on the PDF on the above page on NMbtc's website, you can see detailed graphs of their performance. Notice the relationship between impedance and CFM.

NMB fans tend to have a fairly smooth slope. Most the fans take a big hit in CFM when they hit a certain back pressure point. In the NMB fans this wall is less pronounced than many others. A good flexible CPU fan should have a smooth slope on that graph.

I believe that most good case fans likely have a bigger and earlies drop of off performance against high impedance.

I would like to see Mike find a fan with well documented Specs like the 4710KL-04W-B50 and use it as a reference fan to compare against other CPU fans.

I really wonder, what does that graph look like for the Nexus, the Slipstream, the Sflex and the Noctua? I think knowing that might help us better select certain fans for certain situations.
Which PWM, CPU or Specialty Fan do you Want SPCR to Review?
viewtopic.php?f=9&t=58350&start=30
Last edited by ces on Wed Mar 30, 2011 12:08 pm, edited 1 time in total.

Deucal
Posts: 100
Joined: Sun May 02, 2010 3:10 am
Location: Iceland

Re: What I have Learned about Computer Performance from SPCR

Post by Deucal » Tue Mar 29, 2011 5:07 am

ces wrote: By the way, if anyone thinks this is a result that I am hoping for they would be sadly mistaken. Once Intel succeeds in killing of its competition, it will alter its behavior and we will all lose. Even people who never by AMD... Especially people who never buy AMD.
Intel will never kill off AMD completely, simply because of self preservation. Once Intel becomes a monopoly it's going to get divided up by the US governments anti-trust agencies.

Edit: Kudos for the thread.
Edit2: Typo.

ces
Posts: 3395
Joined: Thu Feb 04, 2010 6:06 pm
Location: US

Re: What I have Learned about Computer Performance from SPCR

Post by ces » Wed Mar 30, 2011 7:33 am

You Can't Always Rely on What You Hear on the "Internets" about Computers (or anything else for that matter)

Before you make a decision or take action, always verify from several sources the assumptions on which you are operating. SPCR is probably the highest quality source of desktop computer information around, but even here the forum postings can come from anwhere. And people can get worked up about the silliest things. To Wit:

Sample 1
colm wrote:smart or stupid, it is an extreme sprite.

the only famous extreme sprites in history are from the threat of death young.

live fast and die.

that is japan.
call it genious, call it what you will..

it is all so grotesquesly stupid, they are just gonna shrink like the land they are loving on. random sprites of desperation and quick change.. like every dainty product we get.

does this deserve a ban? "Ban". where have I heard that word before..
Link to Original Message and thread: viewtopic.php?p=537114#p537114

Makes sense to me :D

Sample 2 (even after a re-edit by the poster)
quest_for_silence wrote:
ces wrote:
tim851 wrote:1.71 dB is the sound pressure rise over ambient noise.
That is what I figured.
Really? What a stupid boy I am, I haven't found the words "rise" or "over" and misinterpreted those data...

...so, in the case of the Antec TrueQuiet, if the ambient noise were around 10dB (a semi-anechoic chamber) the actual sound pressure would be around 12dB, if it's 20dB it would be 22dB, if it were 40dB it would be 42dB... how many dB? Uh, the ambient noise would look like undeclared...

And what about the margin of tenths of dB, apparently with a small, handy instrument? I would want one, it's really impressive (MikeC contents himself with 1dB)...

...or what about the renowned Scythe Slipstream M? It should give 50% more CFM than the Slipstream L, but it would be 40 (forty) times louder than the latter, and about 100 (one hundred) times louder than the above mentioned Antec True Quiet.

Hey ces, look: we have a noise champ!

Not to mention the Gentle Typhoon AP13: at equal CFMs the Slipstream M would be almost 8-10 (eight-to-ten) times louder than it...
Have you ever heard a Slipstream spinning at 1200rpm? And the GT? Not to mention that a 20dB rise over an ambient - say - of 20dB will give 40dB... Have you ever heard a true 21dB level? And a 13dB one? And a 40dB?

...please, I really don't say a word about their proficiency, just there are a lot of things apparently undocumented, and this is rough for logic if you're not a genius (are you a genius?).

(RE-EDITED the message)
Link to Original Message and thread: viewtopic.php?p=537817#p537817

quest_for_silence was a poster whose posts I respected and listened to quite carefully until I saw this. I will still probably listen to him in the future, but probably with more caution than in the past.

People make mistakes. I know I have and I will again in the future. Some people have one point of view and will listen to no other... even when the facts weigh against them. Some people have entire belief systems built around certain products or brands. And you can't always tell.

My rule of thumb is verify 3 times from 3 different sources, then act (sort of a version of the "measure twice and cut once" rule). It cuts down on RMAs

andymcca
Posts: 404
Joined: Mon Jan 11, 2010 8:19 am
Location: Boston, MA, USA

Re: What I have Learned about Computer Performance from SPCR

Post by andymcca » Wed Mar 30, 2011 9:24 am

I actually agree with you that no source can be trusted, even SPCR, but the quote from quest is not a great example. I agree with him that three significant figures of data is meaningless without a methodology. And noise floor is a huge issue when you are talking <2dB, since if the noise floor was 18dBA, that is terrible. Edit: I grant he did go a little wild, though :)

I think the lesson is that every comment should be considered for what it is: a piece of the big picture. Every experience will be different, even between samples, and some people care about things other people overlook. No one is wrong (unless they are Wrong), but no one is right. :D

ces
Posts: 3395
Joined: Thu Feb 04, 2010 6:06 pm
Location: US

Re: What I have Learned about Computer Performance from SPCR

Post by ces » Wed Mar 30, 2011 12:07 pm

ITX Motherboard Brands
ces wrote:If you look at the reviews on Newegg... there seem to be a lot of unhappy motherboard users. Even the best boards have a fair number of unhappy users.

It appears that the top 4 brands by reputation (in no particular orders) are Asus, Gigabyte, Zotac and Intel. Asus seems to be falling. Gigabyte and MSI seem to be rising.

No one seems to like Intel much, but they always seem to come in as the most power efficient. Apparently people have issues with Gigabyte regarding fan control. Myself, I like to just set fans at the speed I want them to run and not waste time futzing with different control settings. Most people here seem to like the futzing.

Zotac seems to often use the most power. And at least until recently (I haven't checked their H67) don't overclock on their ITX boards. Gigabyte does. Asus probably does too.

Gigabyte seems to have the easiest automated overclocking. Asus seems to have the most sophisticated automated overclocking.

Gigabyte makes a lot of noise about having double thick copper on its high end boards... for better cooling. Sounds nice.

It is important to get those famous Japanese solid state capacitors (they are often the first parts to wear out)... but beats me how you quality assure the capacitors their marketing departments are so quick to tout.

That is what I have been able to learn on this subject. The two safest choices are probably Asus and gigabyte. But.. if low power utilization running a Sandy Bridge is your paramount concern. I would choose Intel hands down.
An example of Intel low power utilization with Sandy Bridge. I think this might be a full size board, which makes its power utilization all the more impressive.
jeffy1021 wrote:I just wanted to give an update stating that I just built an i5-2400 machine and it idles at about 30 watts at the wall. This makes sense given the estimated DC wattage in the article. If I got a picoPSU, I could probably save an additional 10 watts.

i5-2400
Intel DH67BL motherboard
OCZ Vertex 2 SSD
Seagate 500GB 2.5" 5400RPM drive
BD/HD-DVD drive
Seasonic SS-300 80 Plus Bronze
Last edited by ces on Wed Mar 30, 2011 1:47 pm, edited 2 times in total.

ces
Posts: 3395
Joined: Thu Feb 04, 2010 6:06 pm
Location: US

Re: What I have Learned about Computer Performance from SPCR

Post by ces » Wed Mar 30, 2011 12:38 pm

andymcca wrote:I actually agree with you that no source can be trusted, even SPCR, but the quote from quest is not a great example. I agree with him that three significant figures of data is meaningless without a methodology. And noise floor is a huge issue when you are talking <2dB, since if the noise floor was 18dBA, that is terrible. Edit: I grant he did go a little wild, though :)
After studying both samples I was actually able to make some sense out of the first example - I think I understood what he was attempting to say. In the second sample, I got the general drift... but when I focused deep I was not able to understand the examples he was using to make his point. That could be because, what I think it was that he was arguing against, was not what anyone was arguing for. A message accompanied by wild gesticulation and exaggeration... makes me more careful about inspecting the message and making sure I understand it. To the extent I understood it, it just didn't make sense to me.

But you need to be cautious even with pretty and appealing messages as well. Maybe especially with those.

ces
Posts: 3395
Joined: Thu Feb 04, 2010 6:06 pm
Location: US

Re: What I have Learned about Computer Performance from SPCR

Post by ces » Wed Mar 30, 2011 8:41 pm

Fan Testing
mdrumt wrote:Don't know if this has been posted anywhere before.
Has anybody else seen this fan review site? Looks pretty new.
http://atreview.net/en/casefan.html
Each fan has a Youtube video "review" with noise / cfm measurements and quality audio recordings!
Hope you SPCR readers find this interesting!
Cheers
See also:
viewtopic.php?f=9&t=61922
ces wrote:Here is an interesting post:

Fan curves: I can make them
viewtopic.php?f=9&t=61225

The 5dBA 120mm Fan (interesting discussion on Fan Theory)
viewtopic.php?f=9&t=61591&



Ask SPCR to review your favorite fan
Which PWM, CPU or Specialty Fan do you Want SPCR to Review?
viewtopic.php?f=9&t=58350

the fan with the highest static pressure
nucc wrote:hi i am new on this site, found this page when I searched for a 120mm computer fans,
im looking for the fan with the highest static pressure, it should be under 25 dba preferrably less
and move at least 28 cfm

i've have looked at some like these:
Silverstone AP121 Air Penetrator - 120mm
Noctua NF-P12-1300 - 120mm
Scythe Gentle Typhoon 120mm - 1150 rpm

is any of these a good choice or is there a better choice?
Those are probably not dab choices. Dorothy Bradbury used to sell some ball bearing 120mm NMB panaflow fans that generated more static pressure at the low sound levels. Seems like she is out of that business. But I think you can still get sleeve bearing versions.

I would select the 1900 rpm Scythe Gentle Typhoon and downvolt it until you get what you are looking for. I think you can downvolt it to under 800 rpm and it is the same blade. If you look at the blade you can see that it is designed to handle high static pressure in a way that the other two fans are not. The downside is that it takes more RPMs for the Gentle Typhoon to generate the same free air CFMs as the other fans.

ces wrote:1. If the above doesn't work, try a set of one of these along with a controller to turn them up for gaming and back down when not gaming. There is just no substitute for Brute Force.

Scythe Gentle Typhoon 120mm x 25mm Fan - 3000 RPM (D1225C12B7AP-29)
http://www.frozencpu.com/products/12363 ... _Soon.html

Scythe Gentle Typhoon 120mm x 25mm Fan - 5400 RPM (D1225C12BBAP-31)
http://www.frozencpu.com/products/12365 ... =g36c15s60

These high speed GTs have an interesting ring to help direct airflow (and keep the blades from flying apart).

2. Here are the silverstone fans that I mentioned:

Silverstone Air Penetrator AP121 120 x 120 x 25mm Fan - UV Blue Blades (SST-AP121)

Take a look at the Youtube demo here:
http://www.frozencpu.com/products/11259 ... =g36c15s60

3. Here is the Nexus 120mm Beamair Fan Gril

http://www.nexustechnologyusa.com/c/ntu ... amair.html
birthdaymonkey wrote:Reporting back: I've got the 1900 Slipstreams installed, and so far I'm pretty impressed. They make quite a bit more noise than the 1850 GTs at full blast, but they also move more air. My motherboard temps were a couple degrees cooler under stress test than they were before, although the CPU temp seemed about the same--I think I've reached the point where my HSF is getting plenty of air, so adding more case airflow the mix doesn't do much.

At idle, when I've got the Slipstreams dialed down with my fan controller, they sound better than the GTs. As I mentioned, one of my GTs had a bit of bearing noise, which isn't an issue with the Slipstreams. GTs also make a unique tonal sound, even at low RPM, which annoyed me a little. The Slipstreams don't make any motor or bearing noise when they're turned down, no clicking or chatter either. I can't tell you what voltage they're running at currently, but they're almost silent except for a faint woosh, and they seem to be providing plenty of flow for playing music/video.

Thus, my conclusions: In the "unique" environment of the Lian Li PC-V351, the Slipstream 1900s might offer better bang for the buck than the GT 1850s. The GTs are much quieter at full power, but the Slipstreams provide better airflow at full power. At low power, the Slipstreams are quieter than the GTs while still providing adequate flow.

I've only had them in for a day, so I haven't completely made up might mind about which fans will stay. Right now I'm leaning toward the Slipstreams. At $8 apiece, I think they have the edge over the twice-as-expensive ($15) GTs for my application.
Shai-tan wrote:Thanks guys. This is the sort of answers I was looking for and I realize my specific case wasn't explained better (it was 3am in the morning here)

Basically I was wondering if there was nothing impeding the fan pointed to the rear exhaust fan should I use a fan with better Air flow vs Less Static pressure on the back of the CPU cooler.

The fans I'm using here are Noctua NF-S12B and Noctua NF-P12.
NF-S12B @ 900rpm
Airflow (m3/h) = 75.8
Static Pressure (mm H2O) = 0.76

Noctua NF-P12 @ 900rpm
Airflow (m3/h) = 63.4
Static Pressure (mm H2O) = 1.21

#1 at the front fan I'm using a NF-P12 because the static pressure is better suited for the impedance between the front and the CPU.
#2 at the CPU cooler I'm using another NF-P12 because the static pressure pushing through the Noctua NH-U12P SE2 would be better.
#3 on the back side of the CPU cooler (facing the rear of the case) I'm using an NF-S12B because there is nothing impeding the fan and the airflow goes straight to the rear fan
#4 Rear exhaust fan is a NF-P12 currently because I still haven't cut out the rear fan cover in the case and the pressure would be better suited for that.

I need to know whether I'd be better using another NF-P12 on #3 or sticking with the better air flow of the NF-S12B

Post Reply