Cheap Socket 754 upgrade

[ElectricLegs]

Originally posted by: rwrentf

"And actually companies do stay in business building crappy products, you yourself proudly proclaim you have the ECS board. My first PC was an 8088 23 years ago, I know hardware. You probably don't even know what an 8088 is NOOB. I've dealt with enough hardware incompatibility hassles from companies that build sub-par products to know it's not worth my time to save the $20. I'll look for a deal, but I'll never use ECS or PC-Chips.

Joke of the day:
"The 939 platform is technically DEAD with the introduction of 940 this year. But... A much better deal is the ECS NF3 MB and the Sempron 3100 for $80 without any rebate."

I've been a constant upgrader since the 8088/386 days too. Many higher tier brands like Asus and Abit have had crap loads of finicky or problem boards. It's not as much the mfg as the actual board or chipset itself. ECS has put out some junk but they've also had some extremely reliable boards like the D6VAA for awesome prices. A smart shopper will wait 6+ months from introduction on any hardware purchase and research a specific model not brand.

 

[Harvey]

Originally posted by: rwrentf
So your saying it's cheaper to build a complete new system than to upgrade a processor or video card after a few years? I've had many systems last me 5+ years upgrading key components.

That's quite often true for several reasons:

1. Technology is advancing so fast that selective upgrading is not always possible. For example, an older AGP 4 motherboard won't be able to take full advantage of an AGP 8 vid card, and a brand new motherboard with an AGP 8 slot won't handle a PCIE card.

RAM is also progressing, and older, non-current devices can cost more than the same amount of faster RAM on a new board. SD-RAM is almost twice as expensive as DDR, and DDR 2 is now going mainstream.

The same is true for ATA 100/133, SATA and SATA 2 drives.

2. An entire older, but still competent, working system can be worth more than the sum of its parts. You can often sell it for enough to cover the difference between upgrading the old setup and the cost of an entire new system, and you end up with two systems, instead of one and some loose parts.

3. The parts for a new setup will be contemporaneous. For the vocabulary challenged, that means, all the parts will be of the same vintage, and the possiblity is much lower that you will enocunter incompatible technologies.

I'm currently running an Athlon 1800+ T'bred (1.5 GHz) o/c'd to 2 GHz (2400+) with a 266 MHz FSB. The motherboard has a VIA KT400 chipset and only one SATA 1 port, which is OK because I have a pair of 80 GB Maxtor ATA 133 drives (one in a moble rack for Ghosting).

It's a great system, and it'll be more than fast enough for a lot of jobs for some time to come so it's worth some money on the hoof. OTOH, if I wanted more performance, which part should I "selectively" upgrade, and what would I do to get any value from the parts I pulled with minimum hassle?

Bottom line -- I've found that upgrading an older system usually is not as cost or performance effective as building a new one.

As for ECS boards, I live in L.A., and we have lots of Fry's stores. If a new board fails (and it has happened), all it takes is a quick local drive to exchange it. You just have to know what you're buying.

 

[furballi]

Who's stopping you from flushing your $ down the toilet? At $90 per combo, I can afford to rebuild every two years with up-to-date components that will run circles around an upgrade rig!

A top tier MB mfr like ECS operates on very slim margin. They make $ on VOLUME. No company will stay in business without turning a profit. And there's no profit stream without repeat orders. That's BUSINESS 101 for the NOOB. Abit was sold recently because their products did not offer sufficient bangs for the bucks.

The board's layout plus bells and whistles are the keys variables among today's MBs.

There are those who will never touch an ECS board because it will not run at +300MHz FSB. So how many AMD-based MBs in service today run at that speed? Probably a fraction of 0.1%. I would guess that less than 5% of the PCs in use today are overclocked. Would a smart company with a sound business plan go after those customers?

This $40 ECS runs everything in my arsenal and is glitch-free up 280MHz FSB (highest Vcore that MB will support). That's VALUE.


So you know about hardware cause you were around when IBM introduced their PCs? Just more proof that you can't teach an old dog new tricks!!!

Toyota and Honda were crap mobiles in the early 70s.

 

[rwrentf]

This old dog know's what's a waste of time and what's not. The extra 5% to 10% speed you get out of your overclocked ECS VALUE board is so worth the 50% MTBF you'll get out of it. Hopefully it won't crash while you're doing your term paper, but maybe it will.
The board's layout plus bells and whistles plus quality capacitors (Like ASUS uses in most of their boards) and good QA (like ASUS uses) are what sells me a board. I know ASUS has put out crap too (they have to compete with the slim margin ECS boards just like all the other manufacturers do), but I won't buy SIS chipsets either. I understand that the profit stream equates to continued business, their just not going to get my cash and waste my time as I return failed board after failed board. I understand I may get one that's stable, but based on the years I've worked on computers I know it's a crap shoot. Have fun with your VALUE board, hopefully the next VALUE board you pick up won't fry your memory or any of your cards.

 

[compcons]

Originally posted by: furballi
The 939 platform is technically DEAD with the introduction of 940 this year. It's a strategy used by AMD and INTEL to force consumers to buy an NEW PC. SLI is a niche market, and will never reach mainstream popularity.

Upgrading processor is a low return on your investment due to the rapid advancement in technology. Ignorant PC builders try to justify the higher component cost with the old "upgradability" line. The best strategy is to build the cheapest rig that will meet your needs today, and possibly the next two years, then build another rig on the 3rd or 4th year.

A few observations:

1) I think you mean Socket M2, not 940. 940 is "old". I'd wager that 939 will host the Semprons, M2 will handle Athlon64's and 940 will continue to manage the Dual processor systems unless M2 takes those honors.

2) The upgrade battle is an interesting one. There cna be some good imemdiate returns when the price is right. In some cases, a new proc will carry someone for another year, making incremental upgrades before the "big ones" effective. I have a customer who has a new gaming rig every 3 years with minor improvements each year (RAM and procs usually). From a cost perspective, he actually has a working model so long as he sticks to it. For some, wholesale upgrades every 2-3 years work because the hardware meets their requirements. Rarely does that strategy work for gaming geeks. It's hard to justify a single $2k expenditure every 2.5 years, but spending $2k across 24 months is ok for them.

Just my thoughts and experiences,
EH

 

[gvbjr]

I believe the combo deal is great although you don't get a heatsink and fan with it and you settle on the mobo. For someone wanting a little more customization I think the 3000 is a fine alternative. Either way you are in the sweet spot as far as bang for the buck. I don't think overclocking yields the same kind of results as it did back in the days of the older technologies, today its mostly a hobby type of thing. However some better yields are definitely available out of these lower priced but in my mind premium products. Another misnomer bandied about is this notion of "dead" motherboard technologies. There will always be a new board coming out with a new format and new options as technology i developed, its just the nature of the beast. I think anyone that has an interest in building these machines would never even consider using a board that long anyhow, so wheres the dilemna.