HP Pavilion a1630n + a1640n: AMD/Intel desktop PCs

Want to talk about one of the articles in SPCR? Here's the forum for you.
MikeC
Site Admin
Posts: 12285
Joined: Sun Aug 11, 2002 3:26 pm
Location: Vancouver, BC, Canada
Contact:

HP Pavilion a1630n + a1640n: AMD/Intel desktop PCs

Post by MikeC » Wed Apr 11, 2007 10:35 am


andyb
Patron of SPCR
Posts: 3307
Joined: Wed Dec 15, 2004 12:00 pm
Location: Essex, England

Post by andyb » Wed Apr 11, 2007 1:23 pm

Im only on page 1 and I can honestly say that I have not read such a sensible comparative review of "real" PC's that "really" get bought by the thousand.

I have no doubt that I will enjoy the rest of this review when I get back.


Andy

PS: I see an awful lot of Packard Bell PC's with faulty PSU's (Bestek), they are total shyte and I would pay money to remove it and put in something that wont toast your motherboard or HDD in 12-months when it goes BANG and starts smoking.

J. Sparrow
Posts: 414
Joined: Wed Jan 17, 2007 7:55 am
Location: EU

Post by J. Sparrow » Wed Apr 11, 2007 1:30 pm

It's irrelevant if you're not obsessed about playing the latest (mostly violent) computer games at the fastest possible speed with the greatest amount of lush, realistic detail.
That sidenote made me wonder 'why didn't they test 3D performance using Frets on Fire ?' :)

continuum
*Lifetime Patron*
Posts: 213
Joined: Tue Jul 19, 2005 10:23 pm
Location: California
Contact:

Post by continuum » Wed Apr 11, 2007 1:34 pm

I would consider pulling both heatsinks and checking for proper thermal paste/pad bonding.

As someone who has worked extensively with Unisys and other onsite repair techs (including in-house techs for several major "appliance" vendors you find in most large server rooms), I would not be surprised if the Intel system, upon repair, never had the CPU thermal interface material changed or even changed properly. The evidence of the digilgence in repair is pretty obvious from the way that the cables inside the system are no longer cable tied together at the proper assembly points.

Also if you were to put the same HD in each system, how do you think that would affect noise/temperature levels?

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

Post by MikeC » Wed Apr 11, 2007 2:04 pm

continuum wrote:I would consider pulling both heatsinks and checking for proper thermal paste/pad bonding.

As someone who has worked extensively with Unisys and other onsite repair techs (including in-house techs for several major "appliance" vendors you find in most large server rooms), I would not be surprised if the Intel system, upon repair, never had the CPU thermal interface material changed or even changed properly. The evidence of the digilgence in repair is pretty obvious from the way that the cables inside the system are no longer cable tied together at the proper assembly points.

Also if you were to put the same HD in each system, how do you think that would affect noise/temperature levels?
Interesting observations. Made me open both up and take another look.

Firstly, the cables on the repaired system are strapped together. Maybe not quite as tightly as in the other system, but close enough. I don't see that there would be any airflow or other functional differences between the two systems due to messy wiring.

Secondly, the Intel HS is a bolt-through type, with the bolts going right in from the top. All 4 bolts were secured all the way, and when I removed the HS, there was a decent layer of TIM between. Slightly sloppy but not excessive. I think the HSF was doing its job fine, and as I mentioned in the review, there was nothing untoward in system behavior even during prolonged CPU stress loading.

Thirdly, I repeated several times that the HDD-related acoustics were just the luck of the draw, and certainly if the HDDs had been swapped, the noise advantage would be with the Intel system. Generally, you cannot specify which model/make drive you want with most systems from big brands -- you can only specify capacity and speed.

Finally, there was no real thermal advantage in either system; both were perfectly fine. All we're really seeing is different thermal diodes reporting on different devices via different motherboards, there's nothing significant we can conclude from the comparison, thermally speaking.

jaganath
Posts: 5085
Joined: Tue Sep 20, 2005 6:55 am
Location: UK

Post by jaganath » Wed Apr 11, 2007 2:37 pm

we swapped out the stock Bestec power supply for a new 350W Seasonic model that's been in the lab for testing.
could you divulge any more titbits about this new Seasonic? is it part of the S12 line or a totally new one? modular cabling?

andyb
Patron of SPCR
Posts: 3307
Joined: Wed Dec 15, 2004 12:00 pm
Location: Essex, England

Post by andyb » Wed Apr 11, 2007 3:31 pm

Having read the rest of the artice, I can say that this is an excelent article, and its nice to see SPCR actually doing comparative benchmarks within games and actually reporting on the gameplay and not just SPCR's usual domain of power, cool, and quiet.

More specifically this review is centered on 2 PC's that people will actually buy (or build with the same/similar components) and not just fantasize about.


Andy

Howard
Posts: 275
Joined: Wed Feb 08, 2006 5:33 pm

Post by Howard » Wed Apr 11, 2007 3:39 pm

It is just me, or do the machines look like total crap?

Atmosper
Posts: 103
Joined: Thu Aug 17, 2006 7:14 am
Location: Västerås, Sweden

Post by Atmosper » Wed Apr 11, 2007 4:28 pm

It is just me, or do the machines look like total crap?
Don't know how you look :wink:, but yes, they do look like they ought to be strangled at dawn.

mattthemuppet
*Lifetime Patron*
Posts: 618
Joined: Mon May 23, 2005 7:05 am
Location: State College, PA

Post by mattthemuppet » Wed Apr 11, 2007 4:37 pm

I'm seriously impressed by the AMD PC's low idle state - that's really getting into the realm of MOTD type power draw. I'm guessing that the older architecture compared with the C2D is the cause of it's higher consumption under load (might also be the chipset too). Was this a 65nm Brisbane part? Makes you wonder what would be possible with a X2 3600, 2.5in HDD and a PicoPSU!

On the noise front, considering that most peoples sensitivity to noise is considerably less than those that read + contribute to SPCR and live in noisier environments, then these PCs most likely do qualify as quiet. If not quiet as we'd class it, then at least unobtrusive. Using rubber grommets for HDD mounting would be a relatively cheap and very effective improvement though, but when margins are so slim and the big boys compete pretty much just on price it'll be a hard sell to the manufacturers.

angelkiller
Posts: 871
Joined: Fri Jan 05, 2007 11:37 am
Location: North Carolina

Post by angelkiller » Wed Apr 11, 2007 5:52 pm

:?

I think this article is missing a key element . I don't think the two systems are comparable. Apples and oranges. The article doesn't mention how much faster the C2D is than the X2 4600. That would explain the difference in power consumption. I don't see how power consumption can be tested if the performance levels isn't the same. Of course a Sempron uses less power than the QX6800. I really hope it does! But the QX680 is several times faster! How can power consumption be tested if a slower processor is paired against a faster one? And since the system came from AMD, I see why they provided a weaker chip. CPU benchmarks would have showed the difference.

I'm not trying to attack anyone, I just wanted to respectively say that I found the article biased to AMD, by leaving out performance comparisons. If I turn out to be wrong, I will gladly correct myself.

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

Post by MikeC » Wed Apr 11, 2007 6:44 pm

angelkiller wrote:I'm not trying to attack anyone, I just wanted to respectively say that I found the article biased to AMD, by leaving out performance comparisons. If I turn out to be wrong, I will gladly correct myself.
uh.... what were the PCMark and 3DMark tests if not performance comparisons? PCMark, especially, tests the entire system.

One very important thing to keep in mind -- the power difference is mostly at IDLE -- when the CPU is doing nothing. What it suggests is that Intel's optimizations for reducing power demand at idle for its mainstream CPUs have not been as successful as AMD's so far.
Last edited by MikeC on Wed Apr 11, 2007 8:08 pm, edited 1 time in total.

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

Post by MikeC » Wed Apr 11, 2007 6:54 pm

jaganath wrote:could you divulge any more titbits about this new Seasonic? is it part of the S12 line or a totally new one? modular cabling?
My guess is that it will become part of the S12 line -- perhaps a new version. The sample is totally OEM-like -- plain-jane. But more efficient than just about any Seasonic (and therefore PSU) before -- except for the 300W SFX 80+, which remains the efficiency champ at low power. (Challenged only by a good AC/DC adapter and the PicoPSU.)

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

Post by MikeC » Wed Apr 11, 2007 8:07 pm

Atmosper wrote:
It is just me, or do the machines look like total crap?
Don't know how you look :wink:, but yes, they do look like they ought to be strangled at dawn.
Wow, do PCs need to pass a beauty pageant or something? :lol:

They're just machines.

The cases and PSU are pretty generic, but that's about par for modest price big brand systems, in my experience. If the HDD wasn't hard bolted, the chassis would probably not vibrate at all.

mattthemuppet
*Lifetime Patron*
Posts: 618
Joined: Mon May 23, 2005 7:05 am
Location: State College, PA

Post by mattthemuppet » Wed Apr 11, 2007 8:09 pm

my only aesthetical objection would be all those bloody stickers, but presumably they can be removed?

Some people actually like their computer to look like a computer, rather than an oversized iPod or undersized fridge :)

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

Post by MikeC » Wed Apr 11, 2007 8:11 pm

mattthemuppet wrote:my only aesthetical objection would be all those bloody stickers, but presumably they can be removed?
Oh yeah, of course. I took them off one & the adhesive came away cleanly with the stickers.

Tzupy
*Lifetime Patron*
Posts: 1561
Joined: Wed Jan 12, 2005 10:47 am
Location: Bucharest, Romania

Post by Tzupy » Thu Apr 12, 2007 3:42 am

@angelkiller: the 2.4 GHz X2 is 10-15% faster than the 1.86 GHz C2D in most applications. Only when heavily using SSE2 code the C2D may come ahead (when using x87 math the X2 is a winner by a large margin). How many office applications are so optimized for SSE2? If they are custom applications, then probably not. From my experience, just enabling SSE2 optimization in the compiler doesn't give a significant performance boost on my C2D E6600, only about 5%.
@MikeC: IMO the main reason of the relative power efficiency of the AMD system is that the G965 chipset is a power hog. It would be nice if SPCR could confirm (or not) my opinion.

angelkiller
Posts: 871
Joined: Fri Jan 05, 2007 11:37 am
Location: North Carolina

Post by angelkiller » Thu Apr 12, 2007 6:14 am

MikeC wrote:uh.... what were the PCMark and 3DMark tests if not performance comparisons? PCMark, especially, tests the entire system.
Tzupy wrote:the 2.4 GHz X2 is 10-15% faster than the 1.86 GHz C2D in most applications. Only when heavily using SSE2 code the C2D may come ahead (when using x87 math the X2 is a winner by a large margin). How many office applications are so optimized for SSE2? If they are custom applications, then probably not. From my experience, just enabling SSE2 optimization in the compiler doesn't give a significant performance boost on my C2D E6600, only about 5%.
OK, I'm wrong. After doing more research, I found that the x2 4600 is in fact marginally faster, and I seem to have missed the PCMark05 scores. :oops: I understand, and now see your view. My bad! Sorry for the disturbance.

cAPSLOCK
Posts: 224
Joined: Fri Sep 16, 2005 11:06 pm
Location: Switzerland

Post by cAPSLOCK » Fri Apr 13, 2007 1:52 am

Was the X2 4600+ a standard version or an EE version? I wonder what kind of "customisations" the Asus mobo has compared to the retail version (M2NPV-MX), I can see that they've scrapped the [useless] PCIe 1x port, the memory slots are all the same colour and closer together, the sata ports are different colours, there are some different chips (for usb or sound or whatever) and the capacitators in the VRM are arranged differently... How much of these minor variations are just different steppings of the motherboard, how much of these changes are asked for specifically by HP and more importantly, what difference does it make?

I don't expect you to answer these questions but they are interesting to think about none-the-less. Perhaps the VRM is tuned for efficiency rather than stable voltages for overclocking... In that case it could be fun trying to find a place to buy the OEM version :twisted:

Update: I just found that the 6150LE integrated gfx runs at a lower clockspeed than standard 6150 (425MHz instead of 475MHz). This obviously contributes to power savings, I wonder what the performance difference is with the 6100 that also runs at 425MHz, probably only different in features (no HD acceleration, DVI support).

tempeteduson
Posts: 189
Joined: Sat Jun 05, 2004 12:56 pm
Location: Houston, Texas
Contact:

Post by tempeteduson » Sun Apr 15, 2007 2:58 am

It uses selected sequences of advanced 3D computer games to accurately gauge the speed of play with those very games.... There's no way you'd want to play any of the games featured in 3DMark with either of these systems in stock configuration.... The advanced games in 3DMark are definitely out of reach for these systems.
Correct me if I'm wrong, but aren't 3DMarks synthetic benchmarks that use proprietary graphics engines not found in "real-world" games? That doesn't make them irrelevant, but worth pointing out.
C&C Generals is... a standout in a long line of RTS games from Entertainment Arts.
I think you meant Electronic Arts, better known simply as EA nowadays. :lol:

Those technicalities aside, I really enjoyed the article. The lower idle (and overall) power consumption of the AMD system is the reason I'm still sticking to their CPUs, and like mattthemuppet said, I'm really interested to see what a 65nm, 1.9GHz 3600+ can do!

Mike, thanks for taking the effort to do practical comparisons like these.

taidi
Posts: 55
Joined: Sun Apr 08, 2007 1:37 am
Location: UK

Post by taidi » Sun Apr 15, 2007 12:25 pm

It's good to see that AMD are keen to exploit their energy efficiency advantage, and high time that PC manufacturers woke up to it.

The energy saving potential of cooler-runnung PCs is too readily dismissed by those who should know better, for example a similar review at Tomshardware,

http://tomshardware.co.uk/2006/09/25/gr ... conclusion

surmised that...

"Were Athlon 64 X2 Energy Efficient processors to reduce average power use by 10 W, the average 10 hour per day office system would cost only few dollars less per year to operate. Multiplying that cost savings over 100 systems, an office manager might save enough to pay for the beverages at the annual Christmas party."

...but this takes no account of the savings to be made in not having to get rid of the heat generated by hot PCs, especially in regions where air conditioning has to be used to maintain a comfortable working environment.

How long I wonder before we see notebook drives fitted as standard in the type of machine reviewed here ? —with a further saving of 8-10W per machine.

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

Post by MikeC » Sun Apr 15, 2007 12:57 pm

taidi wrote:It's good to see that AMD are keen to exploit their energy efficiency advantage, and high time that PC manufacturers woke up to it.

The energy saving potential of cooler-runnung PCs is too readily dismissed by those who should know better, for example a similar review at Tomshardware,

http://tomshardware.co.uk/2006/09/25/gr ... conclusion

surmised that...

"Were Athlon 64 X2 Energy Efficient processors to reduce average power use by 10 W, the average 10 hour per day office system would cost only few dollars less per year to operate. Multiplying that cost savings over 100 systems, an office manager might save enough to pay for the beverages at the annual Christmas party."

...but this takes no account of the savings to be made in not having to get rid of the heat generated by hot PCs, especially in regions where air conditioning has to be used to maintain a comfortable working environment.

How long I wonder before we see notebook drives fitted as standard in the type of machine reviewed here ? —with a further saving of 8-10W per machine.
10W is much too small if the Intel and AMD machines are being compared. It'd be more like >20W according to THG's own data.

2.5" drives for desktops are coming -- they've already been there for servers. Seagate began moving to 2.5" enterprise drives in 2003 and several years ago, Gartner predicted that 2.5-inch enterprise-class drives will be the predominant HDD form factor in multi-user environments in 2007. I don't know 'whether that has happened, but Seagate & Fujitsu have been offering 10k rpm 2.5" enterprise drives for some years; Savvio is Seagate's top enterprise drive series. And Samsung just announced 7200 rpm 2.5" enterprise drives.

Prices for 2.5" notebook drives have been dropping, and capacities are increasing. By year's end, I think we'll see ~500GB at a not horrific price.

Derek Baker
Posts: 264
Joined: Mon Nov 08, 2004 8:05 am
Location: UK

Post by Derek Baker » Fri Apr 20, 2007 7:16 am

MikeC wrote:It's irrelevant if you're not obsessed about playing the latest (mostly violent) computer games...
Is that an axe I hear grinding Mike? :)

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

Post by MikeC » Fri Apr 20, 2007 7:24 am

Derek Baker wrote:
MikeC wrote:It's irrelevant if you're not obsessed about playing the latest (mostly violent) computer games...
Is that an axe I hear grinding Mike? :)
Not really. Just a reflection of the reality, isn't it?

Derek Baker
Posts: 264
Joined: Mon Nov 08, 2004 8:05 am
Location: UK

Post by Derek Baker » Fri Apr 20, 2007 11:08 am

There's reality and whether you like it. How old are you? :)

More seriously, I don't think modern games are necessarily more violent than old ones - they are certainly more detailed in their depiction of it.

jojo4u
Posts: 806
Joined: Sat Dec 14, 2002 7:00 am
Location: Germany

Post by jojo4u » Fri Apr 20, 2007 11:36 am

Well sub-50 W is not too surprising for a hopefully well integrated AMD AM2 system. In this german forum, I can present you some more interesting figures:

AMD X2 3800+ 65nm
MSI K9AGM2-FIH
MDT 1024MB DDR2 PC2-6400
E5-350 BeQuiet PSU
120GB IBM 3,5" hdd
Idle: 34 W

jaganath
Posts: 5085
Joined: Tue Sep 20, 2005 6:55 am
Location: UK

Post by jaganath » Fri Apr 20, 2007 11:41 am

Idle: 34 W
very impressive, esp. w/ a 3.5" hard drive. Is that at the wall (AC) or DC?

jojo4u
Posts: 806
Joined: Sat Dec 14, 2002 7:00 am
Location: Germany

Post by jojo4u » Fri Apr 20, 2007 2:02 pm

jaganath wrote:
Idle: 34 W
very impressive, esp. w/ a 3.5" hard drive. Is that at the wall (AC) or DC?
Wall 220 V. It sounds too good to be true. I'd say it's because of the following factors:
AMD 690G - more efficient than any AM2 chipset.
non-SATA HDD
fairly efficient PSU (test)

Keep in mind that there are other sub-40 W Athlon builds already presented:
viewtopic.php?t=35686&start=0
viewtopic.php?p=268055#268055
viewtopic.php?p=282986#282986
viewtopic.php?p=262839#262839
viewtopic.php?p=272434#272434

stromgald
Posts: 887
Joined: Wed Jun 09, 2004 12:45 pm
Location: California, US

Re: HP Pavilion a1630n + a1640n: AMD/Intel desktop PCs

Post by stromgald » Tue May 01, 2007 12:33 pm

Mike, I just read this article and some things popped out to me as strange. The hard drive temperature for the AMD system was consistently lower, even when the AMD CPU was under load and the Intel CPU was idling (lower temps). Do you think this is just an issue with the HDD thermisitor or is there something else going on here?

I think it might've been better to swap the hard drives and redo a few tests or maybe use the same hard drive for both tests. I know, it's probably tricky with the OS and everything, but it might be more conclusive since the HDD seems to be running hotter and could be driving the wattage up outside of the CPU/chipset combination. Also, did you guys get data on the chipset temps? I always thought nVidia chipsets ran hotter than Intel's, but this points to the contrary. :?

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

Re: HP Pavilion a1630n + a1640n: AMD/Intel desktop PCs

Post by MikeC » Tue May 01, 2007 1:10 pm

stromgald wrote:Mike, I just read this article and some things popped out to me as strange. The hard drive temperature for the AMD system was consistently lower, even when the AMD CPU was under load and the Intel CPU was idling (lower temps). Do you think this is just an issue with the HDD thermisitor or is there something else going on here?
No I don't it was anything odd. They were different drives, different brands.
stromgald wrote:I think it might've been better to swap the hard drives and redo a few tests or maybe use the same hard drive for both tests. I know, it's probably tricky with the OS and everything, but it might be more conclusive since the HDD seems to be running hotter and could be driving the wattage up outside of the CPU/chipset combination. Also, did you guys get data on the chipset temps? I always thought nVidia chipsets ran hotter than Intel's, but this points to the contrary. :?
As you say, this would have been tricky to pull off, and I was simply trying to do a A/B comparison. It has been suggested by someone that EIST on the Intel system could not have been running correctly because they were able to get very similar idle power levels from comparable Intel and AMD systems. I think this is a real possibility but why HP and Intel would not have worked more closely together to ensure better performance is beyond me. As I mentioned in the article, there was no way to control EIST or CnQ in the BIOS of either of these machines, so it would have to have been embedded elsewhere in the firmware -- and both machines had uptodate BIOS when I tested them.

Post Reply