Intel I3-2100 GPU performance, thought it would be fine...

A place to talk about GPUs/Motherboards/CPUs/Cases/Remotes, etc.
erkotz

Posts: 1378
Joined: Mon Aug 22, 2011 9:23 pm
Location:

HTPC Specs: Show details

#21

Post by erkotz » Fri Dec 23, 2011 6:14 pm

FYI, the MCE box I built my parents is based on an I3-2100T (the low-power version of the same CPU) and it handles video fine. Obviously, this is with a Ceton InfiniTV :)
Quality Assurance Manager, Ceton Corporation

winterescape

Posts: 111
Joined: Sun Aug 07, 2011 2:14 pm
Location:

HTPC Specs: Show details

#22

Post by winterescape » Mon Dec 26, 2011 1:26 am

erkotz wrote:FYI, the MCE box I built my parents is based on an I3-2100T (the low-power version of the same CPU) and it handles video fine. Obviously, this is with a Ceton InfiniTV :)
Not sure if you are aware of this but the I3-2100t offers no power savings over the I3-2100. The only reason to purchase this CPU is if you need the low profile heatsink it comes with.

richard1980

Posts: 2623
Joined: Wed Jun 08, 2011 3:15 am
Location:

HTPC Specs: Show details

#23

Post by richard1980 » Mon Dec 26, 2011 1:14 pm

I'm not sure where you got that idea, but it is well known that the suffix "T" on an Intel processor number identifies a more energy efficient model. In addition, you can clearly see from the specifications that the 2100T requires less electricity than the 2100. The 2100T has a lower TDP and a lower Tcase than the 2100.

adam1991

Posts: 2893
Joined: Sat Jun 11, 2011 2:31 pm
Location:

HTPC Specs: Show details

#24

Post by adam1991 » Mon Dec 26, 2011 2:10 pm

...and hence can stand having less of a heatsink?

richard1980

Posts: 2623
Joined: Wed Jun 08, 2011 3:15 am
Location:

HTPC Specs: Show details

#25

Post by richard1980 » Mon Dec 26, 2011 3:24 pm

Yes. The smaller heatsink can be used because the TDP value is lower. But the kicker is looking at the Tcase values. The 2100T has a lower Tcase value, which means it can't run as hot as a 2100. The only way to run cooler with a smaller cooling system is if the processor is more energy efficient. More efficiency = less electricity required.

winterescape

Posts: 111
Joined: Sun Aug 07, 2011 2:14 pm
Location:

HTPC Specs: Show details

#26

Post by winterescape » Mon Dec 26, 2011 5:49 pm

richard1980 wrote:I'm not sure where you got that idea, but it is well known that the suffix "T" on an Intel processor number identifies a more energy efficient model. In addition, you can clearly see from the specifications that the 2100T requires less electricity than the 2100. The 2100T has a lower TDP and a lower Tcase than the 2100.
Sorry, real world testing shows this is a marketing ploy by Intel. Advertise the 2100t to be lower power but overachieve with the 2100 and just charge more for less...

http://www.avsforum.com/avs-vb/showthre ... st20523711

richard1980

Posts: 2623
Joined: Wed Jun 08, 2011 3:15 am
Location:

HTPC Specs: Show details

#27

Post by richard1980 » Mon Dec 26, 2011 6:43 pm

For people that have difficulty reading charts, I have taken the liberty of color coding the information in the chart in the referenced post. The original chart had a column for a 2100 that was underclocked to match a 2100T. However, I find that to be a useless piece of data because that effectively transforms the 2100 into a 2100T. So basically the guy was testing a fake 2100T against a real 2100T. You shouldn't expect any huge difference when measuring power consumption between the two...it's basically 2 of the same CPU at that point.

In the chart below, green means lower power consumption, red means higher power consumption, and yellow means a tie.
Capture.JPG
One thing you'll notice is there is little, if any, difference in the non-CPU intensive tests (idle, BD playback, DVD playback, and Furmark require very little CPU, since of course BD and DVD playback tap the GPU and Furmark is used to test the GPU, not the CPU). The inclusion of the latter 3 of those in the chart really shows the author's ignorance, but that's beside the point. When you look at the things that actually tap the CPU (in this case, the x264 HD benchmark, Prime95, and the Prime95+Furmark tests are the only tests that actually tap the CPU), there is a significant difference in power consumption between the 2100 and 2100T. However, it should be noted that the underclocked 2100 actually scored better in the first two tests, though only by 1 watt. In the Prime95+Furmark test, the underclocked 2100 consumed more power than the 2100T (5 watts more).

In summary, there is a significant difference in power consumption between the 2100 and the 2100T....assuming you actually conduct a real test.

winterescape

Posts: 111
Joined: Sun Aug 07, 2011 2:14 pm
Location:

HTPC Specs: Show details

#28

Post by winterescape » Mon Dec 26, 2011 7:56 pm

I think we will agree to disagree on this one. I look at the same data and draw the conclusion that there is no real world difference, it is in the noise for HTPC tasks.

The author is highly regarded in the HTPC community, I have reviewed his work and I agree with his conclusion, and in addition, the conclusion of the review done by silentpc.

I believe you missed the point on the undervolt/underclock test. This was to demonstrate what you stated. All Intel did was lower the clock and undervolt the same die and charge more for it.

User avatar
newfiend

Posts: 2503
Joined: Tue Jun 07, 2011 12:10 pm
Location: Earth

HTPC Specs: Show details

#29

Post by newfiend » Tue Dec 27, 2011 4:03 pm

sgip2000 wrote:
Scallica wrote:Captain, welcome to the forum.

I would suggest investing in PCie video card. You don't have to spend a lot of money. I have used the Gigabyte GeForce 8400 card in many HTPC systems and it works perfectly. The card can bitstream audio over HDMI and is $30.

http://www.newegg.com/Product/Product.a ... 6814125289
If you want to bitstream HD audio for Blu-Ray, I'd skip that card. It isn't PAP compliant.
A bit better card for bitstreaming would be a EVGA GT430. Had mine over a year w/o issue.
newfiend~

Post Reply