Friday - September 26, 2008 - 6:11AM

New Heatsink Day!
Yesterday I received a new heatsink.  I had been using an Arctic Cooling Freezer 64 Pro.  It was a good heatsink that I had chosen due to my love of the Arctic Cooler Freezer 7 Pro back when I had an Intel E6600.  It performed well, it was inexpensive, it was reliable, and best of all it was quiet at the same time!  It's hard to argue with a combination like that.  The problem is that that particular 92mm air cooler was struggling to keep my leaky AMD Phenom X4 9850 BE under control at load while overclocked.  Temperatures would get as high as 66C when loaded up.  That is most definitely not a good thing.

Eventually, I decided that the time was right to try a new cooler.  I looked around for a while to see what actually performed well for other users--you can't really just *look* at a CPU cooler and tell how good it will be.  Often times, the fanciest looking coolers and the most expensive ones do not sit at the front of the class.  I decided on a Thermalright Ultra 120.  In short order I discovered that it had been discontinued and replaced with the Ultra 120 Extreme model.  "Ok," I thought, "I'll just get that one instead."  Wrong.  It comes in three separate pieces...maybe.  Nobody really knows.  All that anyone knows for sure is that it's probably sold somewhere for some price and doesn't come with a fan but may or may not need extra mounting hardware for AM2 systems.  If it does need mounting hardware, Thermalright helpfully sells three different products for AM2 mounts--all of which warn of the woes you will encounter if you pick the wrong one.

No thanks!  After that confusing mess progressed for a little while, I discovered an offering from Xigmatek.  The Xigmatek HDT-S1283 (catchy, no?) is a 120mm cooler that comes with a high quality fan, a rubber fan mounting system that helps to mechanically decouple the fan from the heatsink (reduce vibration/noise), a fan mount design that leaves a small gap (further reducing fan noise), a three large-heatpipe design, a base that allows the CPU to have direct contact with the heatpipes (woo!), and all the mounting gear needed for either an Intel LGA775 or AMD AM2/AM2+ system.  Performance wise, the charts show it being within testing error of the generally-recognized champion, the Ultra 120.  Assembly was a snap.  Installation was easier than any heatsink I've ever installed.  Best of all, my load temps are now down to about 49C.  No...BEST of all, this thing only costs HALF of what the Ultra 120 does.  The price was about $37 as of this writing.  Newegg has sold it for as cheap as $25 during sales.  What's not to love?  Well, one thing might bug you--if you're on an AM2 system, this may cover your first one or two RAM slots.  It does not prevent you from using them with normal memory modules.  It just might make absurd, tall DIMMs impossible to use or installation and removal of normal DIMMs difficult without removing the cooler.  It has to be mounted horizontally on an AMD system.  Performanace-wise, this doesn't matter.  If you can't handle the non-traditional orientation though, you're SOL on this one.

Anyway, here are the requisite computer pron snaps for you, comparing my previous Freezer 64 Pro with the HDT-S1283:

 

     

Submitted by:  Particle

Wednesday - August 27, 2008 - 3:38PM

So Your Local Screen Dies After Using RDP, Eh?
This is a problem that had been plaguing me for months. It seemed like after a fresh install of Windows Vista or (more recently) Windows Server 2008, RDP would work as expected. I'd login to the machine remotely, dally around, and disconnect. Then when I'd get back to the remote machine in person, I would login and the machine would be as I had left it when using it over RDP. That's how RDP is supposed to work.

Well, RDP doesn't always seem to play by the rules. After some time, the machine I'd remote into would invariably encounter a scenario like this: Log in remotely via RDP(normal), dally around (normal), close the session (normal), go back to the remote machine in person (normal), log into the machine locally (normal), watch the video feed die and no longer have local video (wtf?). At least the machine can still be remoted into and shut down, but that's an inconvenient fix.

As it turns out, the problem appears to be related to Windows and RDP--NOT a particular video driver from ATi or nVidia. The problem is when RDP relinquishes control of the machine (when you successfully login locally) it switches back from its fancy, generic software video driver to whatever the system normally uses for the video hardware. When this happens, your screen will go blank if your monitor isn't connected to your primary video output. Most video cards these days have two DVI connectors. Try switching your video cable to the other one if you are experiencing this problem. No shutdown is required--the cable can be hot swapped and your video *should* magically come back to life.

Just today I did this instead of shutting down the computer via RDP and I'm back up. I don't know if switching the port is a permanent fix like they claim (video output connector # bug) or just the act of unplugging and plugging caused the video card to detect the active display, but it sure beats rebooting.

Thursday - June 19, 2008 - 5:32PM

Online Character Stats Still Busted
For some reason, character stats don't update when the module is compiled as a plugin instead of as an executable. I don't know why yet, but I'll get to the bottom of this one.

Monday - June 09, 2008 - 10:01PM

Character Stats Back Online
Good news, everyone! Online Character Stats are back online. I had fixed the PAT module that takes care of that back in March, but I'd forgotten to update PAT itself with the new revision. D'oh! Anyway, updates will resume like normal starting today.

Bromide Online
More good news. Bromide is back online for the first time since late Feb--his dedicated hardware has been assembled and is now running. It is a Windows Server 2008 Enterprise box with a 3.0GHz 45nm Core 2 Duo (E8400) and 2GB of OCZ DDR2-900 memory. It should stay pretty snappy as its duties increase. This box should be a major improvement over the 1.13GHz Pentium III laptop it used to be on, since that machine had a tendency to shut itself down randomly. Wish it luck!

Tuesday - April 15, 2008 - 8:24AM

A Bit of Progress
Earlier this morning I received an invoice from Newegg for the RMA I have going for the MSI K9A2 Platinum. As is often Newegg's practice, you'll receive an invoice shortly before you get a shipping notice for whatever you buy. If the same holds true for RMAs, I should see the board shipped this afternoon. With a bit of luck, that should mean I will get the new board Friday--just in time for a weekend of fun or failure, depending on how things unfold.

Thursday - April 10, 2008 - 7:55PM

Today is the Day of the Quaddy
It's just too bad that it doesn't matter.  My motherboard still hasn't come back from Newegg, and it probably won't until a week from now.  It could be worse--some RMA departments *cough*Gigabyte*cough* can take a month or longer, so Newegg's two day average turn-around on RMAs is quite refreshing.  Unlike some other companies *cough*Sapphire*cough*, Newegg also doesn't make you pay for return shipping.  Overall, this board failure isn't going to be much of a big deal.  What it does do, however, is stop me from enjoying the silicon sliver below:

Sunday - April 06, 2008 - 12:35PM

Direction of Anger
I did some more investigation on my motherboard that died today. I let the board sit for nearly a day without anything connected to it and entirely without power--not even a CMOS battery. Why would I do that I hear you ask. Well, it might seem crazy, but I've seen it many times before where a seemingly toast motherboard will start to work again after being left as I described earlier. I don't know what causes a board to do that other than possibly an overcharged capacitor, as generally I would think anything preventing a board from working at all would keep it from working again ever.

No joy on my MSI K9A2 Platinum. After sitting alone for a day in a box, my motherboard still refuses to show any sign of life. My brother directed me to this interesting page just about thirty minutes before I retried my motherboard today. This article is curious, as it describes essentially what seemed to happen to my board. It seems that many boards are dying (mostly from manufacturers in the AMD camp) due to high CPU power draw. Three and four phase boards are especially susceptible to this problem. I looked around to see what kind of power my K9A2 board provides. Five phase. Drats. That's not really that far removed from the "oh crap" list. Unlike the three and four phase boards discussed in the article at the link above, my motherboard is listed as compatible even with the 125W TDP Phenom processors at AMD's compatibility site. Even so, to me it looks like all the necessary conditions were apparently present then for my failure to be PWM MOSFET related: five phase power (vs 6 and even 8 phase on some high-end boards), high vCore (1.58V), and high power draw. Estimated CPU draw at load could have been as high as 115 watts. This is still less than the 125W the Phenom could drain in a worst-case scenerio, and this is supposedly supported on my board. Intrigued by this and my brother's find, I took the cooling solution off of my motherboard to get a closer look.



What I found in the cooling solution was both interesting and angering...



The cooling solution seems to have been poorly applied to the PWMs. As you can see, only one row of them are even fully covered by the heat transfer pad. Of the ones that are fully covered, only about half of them had good pressure to the heatsink and pad. I suppose this might explain it--my board could just be a fluke of bad manufacturing.

I guess it's time for an RMA...

Saturday - April 05, 2008 - 10:09PM

A Small Reprieve
While this weekend's topic might be best described as "an epic failure", it seems a small ray of light has emerged from the darkness. The AMD processor that I feared might be dead seems to be fine. I became increasingly suspicious that the CPU might be fine because every other component on the motherboard was still working on my old Intel motherboard--RAM, RAID, GPU, sound...everything. If the CPU had scrambled due to my vCore setting, I would expect the RAM to be damaged. After all, the Athlon 64 X2 processors have a built-in memory controller. The CPU controls the RAM directly...meaning shenanigans would have been likely. I was also suspicious that there was no smell or marks on the processor indicating a physical failure due to runaway thermals.

I guess some times suspicions pay off.

Saturday - April 05, 2008 - 4:38PM

From Irony to Cruelty
Well that was fun... As I posted before, last night was a night filled with silicon dreams. The AMD 5000+ Black Edition CPU that I purchased came in along with its cooler, and I quickly went to work seeing what they could do. After incremental steps and some experimentation, I went to bed with a chip that was apparently stable at 3.1GHz. With some work, a little more juice might be able to be acquired. By noon, however, the system died. It died while operating at 3.30GHz, 1.58V. It died the instant I opened the AMD OverDrive configuration and monitoring utility. Since the box powered down like an over-current event, I didn't think much of it. I planned to set down my vCore a bit due to heat (~58C), but I never got the chance. The system had ceased functioning entirely. Was it a power supply? Nope. Was it the RAM? Nein. Was it an add-in card (video/raid/sound)? Nyet. (Spelled phonetically for your convenience.)

Now I'm back on my Intel Core 2 Duo E6600 that can barely manage 2.40GHz (stock) at 1.5V (not stock...heh). This thing is barely hanging on by a thread, but it's all I've got for a while. At 3.2GHz, the AMD was yielding about 23K MIPS. For comparison, the C2D (back when it actually would do 3.2GHz *rolls eyes*) yielded about 26K MIPS. Yes, the Core 2 architecture is more efficient, but this thing was fixing to clock a lot higher than my aging E6600. I was so close to a faster machine that I could almost taste it.

Sigh.

Saturday - April 05, 2008 - 9:01AM

Updated Guides
I have updated the Guides page to include a new article on how to fix broken desktop icons. I ran into this issue just today, so I thought I would record the knowledge for anyone else that might run into the issue in the future. Searching online for a cure didn't yield anything except for info valid only on Windows 98 and even more stupid or absurd suggestions.

A Bit of Irony
I bought a temporary chip on Thursday to use until the B3 Phenoms list for sale. I wasn't expecting them for at least a week, and in reality probably a lot longer. I even overnighted the chip for $20 so I could play with my new gear this weekend. Later on Thursday, I somehow noticed that the black edition retail boxes lack a CPU cooler. Because of this, I had to place another order for a CPU cooler--also overnight--if I wanted to use the chip I'd paid extra to have here Friday. It was too late to cancel the first order and combine them, so I had to pay another $20 in shipping.

Ironically, the new B3 Phenoms were listed on Newegg starting Friday night. Sure, this chip is ultimately destined to be a gift for a friend so the money would be gone anyway, but I'm just so sick of spending money. Oh well...live and learn.