If someone can make a nice system with Intel for 400 euri, I would like to now. Keep in mind the case must be large (roomy), seems to be the most expensive part.

There are some good inexpensive cases with excellent airflow. I've used the Antec 300 for years. Nice quality with 3 speed 140mm top fan and 3 speed 120mm rear fan. Add a 120mm side fan and 1 or two 120mm front fans and you're set. There are also some minor variations of the basic 300 case now. Other low cost options are the NZXT Source 120 and the Corsair Carbide 200R, both of which will hold enough fans to made a wind tunnel jealous. Have used dozens of the Antec 300 and am in the process of testing the other two. For all 3 the quality is decent. A super expensive case won't get you much more in practical terms. Definitely prefer cases with the PS on the bottom.

Yeah your info would be great as you have build a lot. I never build a system from scratch did a lot of changes, additions, the upgrading bit.
Looked at a lot of cases and want the PSU in the bottom, and one big or two smaller in the front to suck air and one in the back to blow out air, and one or two top fans to blow out air.
Now I see plenty cases with fans in the front, but the front is closed, like the Corsair Obsidian 550D for example. That seems a roomy case, price in Germany acceptable for it. But why is the front closed while it has fans there?
I work in physics but can not explain this yet. There must be side openings for sure, but if that sucks enough air? Its not funneled like a jet engine.
Would be glad for an explanation.

---

Greetings from TJ

I have browsed a bit and searched only in the Netherlands (where a lot of parts are more expensive than in the US). An i7-4770K (four core), 84W, €317. An i7-3930K (six core), 130W, €530.

I don't recommend the i7-3930K, even though it has six cores, it's outdated. If you want to build a high-end cruncher for CPU tasks, you should wait a little more for the release of the 4th generation extreme CPU series. There are going to be models with 8 cores and HT.
If you want to build cruncher PC now, I recommend an i7-4770 or i7-4770K. If this cruncher should have 2 GPUs, I recommend a motherboard with 4 PCIe x16 slots (such as my Gigabyte GA-Z87X-OC). You can put 2 GPUs in a motherboard like that 4 slots apart, so there will be 2 slot space for the airflow when using standard dual slot GPUs, and if you choose some oversized 3 slot wide GPUs, there will be still a slot space between the cards.
Besides being an Intel fan, I prefer ASUS and Gigabyte parts, especially motherboards. I've searched for days for a new motherboard, and I've found that this Gigabyte GA-Z87X-OC is the ideal choice for a dual-GPU cruncher (24/7/365) for the following reasons:
1. price (it's around 180€)
2. PCIe slot arrangement
3. onboard CPU power sulpply effectivity, and long lasting capacitors
4. Thermal conductivity of the motherboard (it has doubled thickness copper in the PCB - you can feel it by its weight)
5. onboard switches and BIOS diagnostic display
This motherboard has one drawback: the PCIe slots share the CPU's 16 PCIe lanes, so in a dual GPU setup both GPUs will run at PCIe x8, which doesn't impair the performance of the current GPUGrid client. However, you can have a motherboard without that limitation for doubled price (e.g. the GA-X58-OC Force)

I'd consider 180€ far too much for a mainboard, especially if it shall "only" run 2 GPUs. In this case I'd choose 2 standard PCIe slots, so 8x PCIe 3 for plenty of bandwidth. And considering TJs current thermal problems I wouldn't recommend he'd try 4 GPUs ;)

And the Ivy-E for socket are rumored to top out at 6 cores + HT again for "enthusiasts". Xeons will go higher (again), but at insane prices for BOINC. Still I also wouldn't recommend the 3930K as it's a 32 nm Sandy Bridge and henceforth less power efficient. Sure, it's got 50% more cores than Haswell, but clock for clock Haswell should be around 15% faster, which combined with the higher clocks reduces the performance advantage further to <30%.

MrS

---

Scanning for our furry friends since Jan 2002

Four GPU's is not to bad if you take two GTRX690 :-)

Sorry for the Intel fans, but I will build an AMD based one first, with the 2 GRX660. I will keep it cheap as I am not that enthusiastic about there performance, but they are here no and must do GG.
Then with Zoltan's tip in mind, I wait for the 4th generation extreme one's to see how to do and what the prices air. That can become a "super cruncher" with two GTX770, or one 690. And would be something for the winter. First for the money reason and second because summer is start to run and then my attic becomes to warm. In fact it already is.

---

Greetings from TJ

There are some good inexpensive cases with excellent airflow. I've used the Antec 300 for years. Nice quality with 3 speed 140mm top fan and 3 speed 120mm rear fan. Add a 120mm side fan and 1 or two 120mm front fans and you're set. There are also some minor variations of the basic 300 case now. Other low cost options are the NZXT Source 120 and the Corsair Carbide 200R, both of which will hold enough fans to made a wind tunnel jealous. Have used dozens of the Antec 300 and am in the process of testing the other two. For all 3 the quality is decent. A super expensive case won't get you much more in practical terms. Definitely prefer cases with the PS on the bottom.

Now I see plenty cases with fans in the front, but the front is closed, like the Corsair Obsidian 550D for example. That seems a roomy case, price in Germany acceptable for it. But why is the front closed while it has fans there?
I work in physics but can not explain this yet. There must be side openings for sure, but if that sucks enough air? Its not funneled like a jet engine. Would be glad for an explanation.

I agree. The Corsair and NZXT above collect ait from the edges of the front panel. That has to be restrictive compared to the Antec 300 which pulls air directly in through the front panel.

As far as the AMD vs Intel question. The 990FX MB you bought has 2 x16 slots, much preferable to a board with 2 x8 or an x16 and x4 slot. I think you made a good choice, I think you'll be happy with the FX-8350 too. Realistically we're not talking much CPU performance difference at all compared to the fastest Intels. People get all excited over 10-15%. Big deal. Those are game benchmarks, not the massively threaded scientific apps we run. Remember too that Intel "owns" the benchmark comapnies and a good share of the "performance" sites, not to mention their dirty compiler tricks. How many CPU BOINC credits is that? Not much. If you use that very significant cost savings to buy faster GPUs you will be so much further ahead with the FX-8350 system you bought, far ahead in total production. Sure you can buy cheap Intel CPUs and MB to get the cost closer, but then the AMD system is faster for CPU too. I used to run more Intel than AMD in the P2 and P3 days but comparing the two, my AMDs always ended up faring better and costing less. As I mentioned I've been building PCs for many years. One thing I've noticed is that I keep having to fix the Intel based systems and the AMDs just keep running and running. I've still got quite a few in businesses chugging along doing their more than decade old accounting and other business chores without me ever having to do more than blow the dust out every few years. Works for me. Again, as I mentioned before what's possibly more cost effective than buying one of the excellent X6 T1045 CPUs for $80 and dropping it into the latest 990FX AM3+ motherboard of choice?

I'd consider 180€ far too much for a mainboard, especially if it shall "only" run 2 GPUs. In this case I'd choose 2 standard PCIe slots, so 8x PCIe 3 for plenty of bandwidth. And considering TJs current thermal problems I wouldn't recommend he'd try 4 GPUs ;)

The mainboard could support 4 GPUs, besides I wouldn't put a 300€ Core i7-4770K in a cheap mainboard. However this mainboard costs 50€ more than I usually spend on a mainboard, but I thought it's worth the extra cost.

As far as the AMD vs Intel question. The 990FX MB you bought has 2 x16 slots, much preferable to a board with 2 x8 or an x16 and x4 slot. I think you made a good choice, I think you'll be happy with the FX-8350 too. Realistically we're not talking much CPU performance difference at all compared to the fastest Intels. People get all excited over 10-15%. Big deal. Those are game benchmarks, not the massively threaded scientific apps we run. Remember too that Intel "owns" the benchmark comapnies and a good share of the "performance" sites, not to mention their dirty compiler tricks.

As far as I know these dirty tricks are not present in the latest versions of that compiler. However, it came to my mind, that the Yoyo client could be "optimized" to AMD CPUs, still we don't call this "dirty tricks". NVidia also used "dirty tricks" for better benchmark results. There are many little things which made me not to like AMD, for example their driver support is worse than Intel's.

I used to run more Intel than AMD in the P2 and P3 days but comparing the two, my AMDs always ended up faring better and costing less. As I mentioned I've been building PCs for many years. One thing I've noticed is that I keep having to fix the Intel based systems and the AMDs just keep running and running. I've still got quite a few in businesses chugging along doing their more than decade old accounting and other business chores without me ever having to do more than blow the dust out every few years. Works for me.

It's interesting, because I had such difficulties with AMD configurations at these times (leaking and bloated capacitors, way too loud cooling), so I gave up on AMD. One of our customer had a PIII 733MHz configuration with 64MB SDRAM, running Win'98(!) until recently, when their ERP system stopped working on Win'98 after an update (mainly for security reasons). A local "wiseman" tried to install WinXP on that PC, but he gave up because it was way too slow, but it's still working. They have an Athlon XP based PC (from a different source) which still working, even though it's hot. We sold many P4 2.8GHz PCs with Intel mainboards and they are still working today. I wish they only fail soon, because when this happens, I don't have to explain why should they buy fresh hardware.

The quad with the 550Ti problems, is more then 5 years old. Its a Dell XPS, small, shiny black edition with a rubber mat on top. The model was only short available. I got it due to an action of dell with 20% discount and some free stuff. It came out the box, ran for 4 year 24/7 without problem, then the GPU broke (stripes on screen) and then I bought the GTX550Ti due to it light PSU and could do GG. It worked fine until yesterday when I swapped the 55oTo for the 660 to test.
But 5 year ago there was a lot of complaining about Vista, it was unstable, nod drivers and so on. But I have never had problems.

Around 15 year back there was a "privet PC project", one could use salary before tax and the employer paid also a part. Then I was the only one who bought an AMD, it was way cheaper, and I had the biggest HD and a CD-burner as well and a 18.1 inch flat screen. There was a limit. So by choosing AMD I had the "biggest system" of all participants. But I would get trouble said IT and colleagues. But also this system server me for 10 years (no crunching).

So it is a bit one own experience and likings. I don´t mind to have all sorts of PC´s

---

Greetings from TJ

If someone can make a nice system with Intel for 400 euri, I would like to now. Keep in mind the case must be large (roomy), seems to be the most expensive part.

There are some good inexpensive cases with excellent airflow. I've used the Antec 300 for years. Nice quality with 3 speed 140mm top fan and 3 speed 120mm rear fan. Add a 120mm side fan and 1 or two 120mm front fans and you're set. There are also some minor variations of the basic 300 case now. Other low cost options are the NZXT Source 120 and the Corsair Carbide 200R, both of which will hold enough fans to made a wind tunnel jealous. Have used dozens of the Antec 300 and am in the process of testing the other two. For all 3 the quality is decent. A super expensive case won't get you much more in practical terms. Definitely prefer cases with the PS on the bottom.

Indeed nice systems Beyond, but after careful examination of pictures I saw one thing that I don´t like. I want the PSU in the bottom, but I want holes in the bottom as well because I want its fan facing downwards, not into the case, where warm (lukewarm) air flows towards the MOBO.
Air in from the side will work I guess, but I like one big or two fans behind a grit (with dust filter).

---

Greetings from TJ

As far as the AMD vs Intel question. The 990FX MB you bought has 2 x16 slots, much preferable to a board with 2 x8 or an x16 and x4 slot. I think you made a good choice, I think you'll be happy with the FX-8350 too. Realistically we're not talking much CPU performance difference at all compared to the fastest Intels. People get all excited over 10-15%. Big deal. Those are game benchmarks, not the massively threaded scientific apps we run. Remember too that Intel "owns" the benchmark comapnies and a good share of the "performance" sites, not to mention their dirty compiler tricks.

As far as I know these dirty tricks are not present in the latest versions of that compiler. However, it came to my mind, that the Yoyo client could be "optimized" to AMD CPUs, still we don't call this "dirty tricks". NVidia also used "dirty tricks" for better benchmark results.

Yoyo's ecm client is simply wrapped and why would ecm optimize for AMD when there are far more Intels running the project?

There are many little things which made me not to like AMD, for example their driver support is worse than Intel's.

What drivers? CPU???

I used to run more Intel than AMD in the P2 and P3 days but comparing the two, my AMDs always ended up faring better and costing less. As I mentioned I've been building PCs for many years. One thing I've noticed is that I keep having to fix the Intel based systems and the AMDs just keep running and running. I've still got quite a few in businesses chugging along doing their more than decade old accounting and other business chores without me ever having to do more than blow the dust out every few years. Works for me.

It's interesting, because I had such difficulties with AMD configurations at these times (leaking and bloated capacitors, way too loud cooling), so I gave up on AMD. One of our customer had a PIII 733MHz configuration with 64MB SDRAM, running Win'98(!) until recently, when their ERP system stopped working on Win'98 after an update (mainly for security reasons). A local "wiseman" tried to install WinXP on that PC, but he gave up because it was way too slow, but it's still working. They have an Athlon XP based PC (from a different source) which still working, even though it's hot.

Surely you know that the bad capacitors had nothing to do with CPUs and were equally common on all motherboards, a result of the motherboard companies buying crappy, defective Chinese capacitors. It was a dark period for motherboards, many failed. A friend and the largest cruncher on our team during that period all Intel with ECS MBs (22 PCs). They all failed (Chinese capacitors), he was not happy. Besides my business experiences, I'll relate a personal one from quite a while ago. I had an equal number of regular AMD desktop systems and very expensive large cache Intel servers spread around my house, one of each in 5 rooms to even out the heat distribution. Lightning storm: everything went dark. When the power returned 3 of the expensive Intel servers were dead, 2 MBs and 1 CPU. Within a week the other 2 failed. All of the AMD systems survived and never showed a sign of the catastrophe. Not a large statistical sample but telling. Pretty much the end of my trusting expensive Intel servers...

Indeed nice systems Beyond, but after careful examination of pictures I saw one thing that I don´t like. I want the PSU in the bottom, but I want holes in the bottom as well because I want its fan facing downwards, not into the case, where warm (lukewarm) air flows towards the MOBO.
Air in from the side will work I guess, but I like one big or two fans behind a grit (with dust filter).

I always install the PS with the fan on the bottom. There's enough of an airspace. Even if you turned it upside down the fan inside the PS would be sucking in the air and expelling it out the back of the case. Holes in the botton are OK, maybe more airflow but the floor level is also where the dirt/dust is. Take your pick, I've heard both sides argued. Big surprise...

Indeed nice systems Beyond, but after careful examination of pictures I saw one thing that I don´t like. I want the PSU in the bottom, but I want holes in the bottom as well because I want its fan facing downwards, not into the case, where warm (lukewarm) air flows towards the MOBO.
Air in from the side will work I guess, but I like one big or two fans behind a grit (with dust filter).

I always install the PS with the fan on the bottom. There's enough of an airspace. Even if you turned it upside down the fan inside the PS would be sucking in the air and expelling it out the back of the case. Holes in the botton are OK, maybe more airflow but the floor level is also where the dirt/dust is. Take your pick, I've heard both sides argued. Big surprise...

I have all systems on a very smooth riser (self made) where dust is easily blown away, it can´t stick. I clean that regularly. Therefor I like the dust filters. But I will see.

...servers spread around my house, one of each in 5 rooms to even out the heat distribution...

Expensive heater. How could you sleep? I also want PC´s on the floor where we sleep, to get my attic cooler, but the misses won´t... to much noise.

---

Greetings from TJ

...servers spread around my house, one of each in 5 rooms to even out the heat distribution...
Expensive heater. How could you sleep? I also want PC´s on the floor where we sleep, to get my attic cooler, but the misses won´t... to much noise.

White noise. Sleeping is better, it hides much of the backgound sounds. Not so expensive, my house is very efficient and uses a type of off peak that has a fixed rate. Summer is not as good, as air conditioning doesn't work as well. This is Minnesota though so heating is a MUCH bigger issue than cooling :-)

I always install the PS with the fan on the bottom. There's enough of an airspace. Even if you turned it upside down the fan inside the PS would be sucking in the air and expelling it out the back of the case. Holes in the botton are OK, maybe more airflow but the floor level is also where the dirt/dust is. Take your pick, I've heard both sides argued. Big surprise...

After reading it again I understand. Off course to fan in the PSU sucks as well, so it could help as extra cooling faced up.

---

Greetings from TJ

I don't understand the PSU at the bottom trend. PSUs suck air from inside the case and thus contribute to expelling warm air. So, why would anyone put them at the bottom, where cool air is?

There's only one reason for me: better cooling for the PSU itself. Maybe important for high-end 1KW PSUs, but for more modest ones, I don't think it makes much sense.

---

Yoyo's ecm client is simply wrapped and why would ecm optimize for AMD when there are far more Intels running the project?

:) Good point. Then their application is simply better on AMD processors. The question is why don't they make a better (optimized) application for Intel processors, if that CPUs are the majority?

There are many little things which made me not to like AMD, for example their driver support is worse than Intel's.

What drivers? CPU???

No. :) I mean the drivers for the chipset around it. I recall some bad VIA drivers causing various problems, or the chipset was way slower than the competing Intels. There was a point when AMD decided to make chipsets for their own processors, because they weren't satisfied with the 3rd party chipsets. Intel did the same for the same reason, but earlier. NVidia chipsets were quite good, as far as I can recall them.

Surely you know that the bad capacitors had nothing to do with CPUs and were equally common on all motherboards, a result of the motherboard companies buying crappy, defective Chinese capacitors. It was a dark period for motherboards, many failed.

Yes, I know. The "defect" caused by a stolen formula of the electrolyte, which turned out to be incomplete at the time it was stolen. So the electrolyte was degraded by normal usage, the result of this degradation was a gas (hydrogen, if I remember it correctly), and that bloated them, reduced their capacity, forcing the FETs swiching more, causing more heat. This process, however, was faster at higher temperatures, so the hotter AMDs failed earlier. The problem had a "financial" component: those, who bought AMD for their lower prices, usually bought a cheaper mainboard for it, and these cheaper mainboards were made of cheaper components. But there were quite expensive mainboards at that time with bloated capacitors. One of my spare board is an Intel D975XBX, which had two of them (one at the northbridge, and one at the DIMM sockets). After I've changed these two to new ones, the MB's stability problems were gone. We had one or two Intel entry level server board failures (after 4 or 5 years of use) because the leaking electrolyte corroded the PCB to an unrepairable condition. I used to desolder the good capacitors (and sometimes the FETs also) from dead MBs since then, to have spare parts for fixing others.

A friend and the largest cruncher on our team during that period all Intel with ECS MBs (22 PCs). They all failed (Chinese capacitors), he was not happy.

The name ECS is sounds good, but their MBs are... not that good quality as the company name suggests.

Besides my business experiences, I'll relate a personal one from quite a while ago. I had an equal number of regular AMD desktop systems and very expensive large cache Intel servers spread around my house, one of each in 5 rooms to even out the heat distribution. Lightning storm: everything went dark. When the power returned 3 of the expensive Intel servers were dead, 2 MBs and 1 CPU. Within a week the other 2 failed. All of the AMD systems survived and never showed a sign of the catastrophe. Not a large statistical sample but telling. Pretty much the end of my trusting expensive Intel servers...

I can understand that. The PSUs in those servers weren't up to their tasks... regarding spike suppression. These spike suppressors could withstand only a limited number of spikes, depending on the power of the spikes they receive, so it is highly recommended to have power strips with built-in spike (or surge) suppressors.

I don't understand the PSU at the bottom trend. PSUs suck air from inside the case and thus contribute to expelling warm air. So, why would anyone put them at the bottom, where cool air is?

There's only one reason for me: better cooling for the PSU itself.

You are right about that.

Maybe important for high-end 1KW PSUs, but for more modest ones, I don't think it makes much sense.

It does for all PSUs, since all PSUs' main electric characteristics are better at lower temperatures. (efficiency, ripple, noise, voltage stability, longevity)

Yes PSU are an important part of a PC, but are mainly forgotten or cheap. The old T7400 (7 years old) has already a 80+ gold edition.

---

Greetings from TJ

I don't understand the PSU at the bottom trend. PSUs suck air from inside the case and thus contribute to expelling warm air. So, why would anyone put them at the bottom, where cool air is?

There's only one reason for me: better cooling for the PSU itself.

You are right about that.

Two other reasons, the heaviest component is then at the bottom of the case and most important, the CPU in a case with bottom mounted PS ends up in the upper back corner with both the back and top fans expelling the hot air. Hot air rises ;-)
3rd reason: Also puts the GPUs in a better position for cooling from the front and side fans.
4th reason: PS efficiency generally goes down as temperature rises (not to mention failure rate).

No. :) I mean the drivers for the chipset around it. I recall some bad VIA drivers causing various problems, or the chipset was way slower than the competing Intels. There was a point when AMD decided to make chipsets for their own processors, because they weren't satisfied with the 3rd party chipsets. Intel did the same for the same reason, but earlier. NVidia chipsets were quite good, as far as I can recall them.

A long time ago and exactly why AMD started making their own chipsets. No issues since that I know of. Intel had the same problems for the same reasons.

Surely you know that the bad capacitors had nothing to do with CPUs and were equally common on all motherboards, a result of the motherboard companies buying crappy, defective Chinese capacitors. It was a dark period for motherboards, many failed.

Yes, I know. The "defect" caused by a stolen formula of the electrolyte, which turned out to be incomplete at the time it was stolen. So the electrolyte was degraded by normal usage, the result of this degradation was a gas (hydrogen, if I remember it correctly), and that bloated them, reduced their capacity, forcing the FETs swiching more, causing more heat. This process, however, was faster at higher temperatures, so the hotter AMDs failed earlier.

I think the P4s were hotter than any AMD yet.

The problem had a "financial" component: those, who bought AMD for their lower prices, usually bought a cheaper mainboard for it, and these cheaper mainboards were made of cheaper components.

If true (it wasn't for me) it's not exactly fair to compare low quality boards to high quality boards. So are you saying that AMD isn't good because you bought poor quality motherboards for them and those boards failed at a higher rate?

Besides my business experiences, I'll relate a personal one from quite a while ago. I had an equal number of regular AMD desktop systems and very expensive large cache Intel servers spread around my house, one of each in 5 rooms to even out the heat distribution. Lightning storm: everything went dark. When the power returned 3 of the expensive Intel servers were dead, 2 MBs and 1 CPU. Within a week the other 2 failed. All of the AMD systems survived and never showed a sign of the catastrophe. Not a large statistical sample but telling. Pretty much the end of my trusting expensive Intel servers...

I can understand that. The PSUs in those servers weren't up to their tasks... regarding spike suppression. These spike suppressors could withstand only a limited number of spikes, depending on the power of the spikes they receive, so it is highly recommended to have power strips with built-in spike (or surge) suppressors.

They all were on APC 600 sine wave UPSes (I bought a bulk purchase of them) and used them both at businesses and here. A couple of the AMD systems were just on surge strips. They weren't damaged and made me wonder about the importance of UPSes...