|
When using this resource, please read FragKrag's opening post. The Tech Support forum regulars have helped create countless of desktop systems without any compensation. The least you can do is provide all of the information required for them to help you properly. |
You could now try to choose a shop to buy from. There's geizhals.de to check for the cheapest prices and shops.
If you buy all parts at the same place, you'll probably save on shipping costs. Mindfactory, which you mentioned, has 0 € shipping if you order after midnight if the price is high enough. Make sure to delete the "gold service" for 10 € (?) from the shopping cart if you buy from them. The other strange thing about them is, their prices often change multiple times a day. I'm not sure what's up with that, if you can somehow use that to save money.
After you decide on a shop, you could try to ask again about parts. There could be special prices in that shop for something that's not exactly what you want, like a cheap 1866 MHz RAM kit. Perhaps they sell a bundle of Z77 board with 3570K CPU for the same price as the parts you want, and modest overclocking would only mean 20 EUR more for a cooler, etc.
|
On April 14 2013 04:14 Myrmidon wrote:
How often do you write large files to the SSD and need that to be fast? (use here might be video capture, extracting files if not zipped or otherwise CPU not the bottleneck; maybe certain server / pro work) Like I said, it's mostly just sequential write where 840 is slower. Not anything noticeable or notably different for most people usually.
Hm, well I do use FRAPS quite a bit, and that would be nice to record to the SSD as well of course. So that would be 20-30 GB sequential (I assume) writes. Sounds like this is where the 840 Pro 128GB would be better suited? I have been writing this to my Samsung Spinpoint F3 1TB HDD before and it's been working reasonably well already though - so no reason to spend the 120 EUR when I could go with the cheap 85 EUR instead. Still unsure.
On April 14 2013 04:14 Myrmidon wrote:
If RAM cheaper than Vengeance is available, you can get that instead.
Other 2x4GB 1600MHz DDR3 RAM (which is what I was looking for, not sure how important 1600Mhz vs 1333MHz is) that is cheaper than the Vengeance RAM is the Corsair PC1600. That's 10 bucks cheaper, so I'll go for the Corsair RAM instead unless somebody has any hint as to which one is considerably better/faster?
On April 14 2013 04:14 Myrmidon wrote:
If you want to save, you can probably use B75 instead of H77, unless you'd really get a second SSD maybe.
Good call, I'm not interested in using 2 SSDs so that's another 5-10 bucks! Thanks!
On April 14 2013 04:14 Myrmidon wrote:
Did you decide on where you're ordering from and the power supply?
I was looking at the parts first, referencing Amazon prices, then looking at some of the specialized dealers (such as mindfactory) for who makes the cheapest offer! As far as the PSU is concerned, I'm currently looking into how much power my components need and what the cheapest, yet still 'okay' PSU is that I can get! Any insight would be helpful here as well!
Thank you again for the support so far!
EDIT @Ropid: Makes sense, that's the next step! Comparing prices!
|
Just FRAPs to hard drive as before.
1600 MHz is only a few percent or less faster than 1333 MHz for most things, sometimes closer to zero. You can just get the cheaper 1600 MHz RAM, or 1333 MHz if it's significantly less.
You can go with Seasonic G360 (SSR-360GP). Less good but maybe cheaper would be Super Flower Amazon 450W, XFX Core (Pro) 450W. There are a lot of other options. I wouldn't recommend spending less, though there are some lower-quality parts that are acceptable and a little bit cheaper.
|
Mindfactory is really really good. Best prices in EU imo (pity outside from Germany it's generally no worth using cos case comes from separate warehouse so you have 30 euro shipping by 2). Boxed 3570k for €188 I've seen a few times on offer. Tat's absolutely insane for EU prices. Hardwareversand is very good too. Another German site, tends to be only a little bit more expensive, but a flat 19 euro shipping rate is very good.
|
On April 14 2013 03:30 Ropid wrote:Show nested quote +On April 14 2013 02:57 NihiLStarcraft wrote:On April 14 2013 02:38 Ropid wrote:
You can get a Samsung 840 Pro for 115 EUR or something like that. The Samsung 840 (without "Pro") is 145 EUR for 250 GB.
Comparing the OCZ Vertex4 with the Samsung 840 Pro, it looks like the Vertex4 is a bit faster both when it comes to reading and writing. And the 840 Pro is only 5-10 bucks cheaper! Not sure if it's worth it? That's not true. Anandtech does benchmarks where they try to simulate what happens in the real world, and the Samsung beats it handily there. I think the Samsung is actually the very best drive. I'm a bit suspicious of OCZ. They had a history of problems regarding reliability. They also have to buy the controller and memory they use, while Samsung can make everything themselves. That's where the better price can come from. It does not have to mean worse quality.
Yeah, and even now OCZ has 5% defects versus 0.48% for Samsung for SSDs.
|
On April 14 2013 07:56 MrCon wrote:Show nested quote +On April 14 2013 03:30 Ropid wrote:On April 14 2013 02:57 NihiLStarcraft wrote:On April 14 2013 02:38 Ropid wrote:
You can get a Samsung 840 Pro for 115 EUR or something like that. The Samsung 840 (without "Pro") is 145 EUR for 250 GB.
Comparing the OCZ Vertex4 with the Samsung 840 Pro, it looks like the Vertex4 is a bit faster both when it comes to reading and writing. And the 840 Pro is only 5-10 bucks cheaper! Not sure if it's worth it? That's not true. Anandtech does benchmarks where they try to simulate what happens in the real world, and the Samsung beats it handily there. I think the Samsung is actually the very best drive. I'm a bit suspicious of OCZ. They had a history of problems regarding reliability. They also have to buy the controller and memory they use, while Samsung can make everything themselves. That's where the better price can come from. It does not have to mean worse quality. Yeah, and even now OCZ has 5% defects versus 0.48% for Samsung for SSDs.
Heh wasn't there some hubbub last year about people misinterpreting 'return' rates as failure rates 
http://www.hardware.fr/articles/881-7/ssd.html
Tom's, of all places, actually voiced the same concern I had about how to interpret this kind of data in their 2011 column.
http://www.tomshardware.com/reviews/ssd-reliability-failure-rate,2923-3.html
+ Show Spoiler +A drive failure implies the device is no longer functioning. However, returns can occur for a multitude of reasons. This presents a challenge because we don’t have any additional information on the returned drives—were they dead-on-arrival, did they stop working over time, or was there simply an incompatibility that prevented the customer from using the SSD?
This information only presents us with more questions. If online purchases account for the majority of hard drives sold, poor packaging and carrier mishandling can have a real effect on return rates. Furthermore, we also have no way of normalizing how customers used these drives. The large variance in hard drive return rates underlines this problem. For example, the Seagate Barracuda LP rises from 2.1% to 4.1%, while the Western Digital Caviar Green WD10EARS drops from 2.4% to 1.2%.
Alright, so the data really tells us nothing about reliability. What does it say then? Well, there seem to be more satisfied French customers purchasing Intel SSDs and not returning them than any other brand. Interestingly, in private chats, engineers from several SSD manufacturers note that a significant percentage of returned drives are unused and sealed, even though customers claim a compatibility issue. Customer satisfaction is interesting, but it's far less interesting to us than failure rates. Moving on.
To be fair, for the more recent Oct 25, 2012 numbers, Hardware.fr does state that there is at least some kind of cross-checking (type de panne, tests croisés) to make that determination. It is rather odd that while they have access to this one French retailer's database they continue to make no effort at presenting us this data in its entirety. I would assume there's some understanding that they won't b/cc manufacturers are bound to get mad, but I would like to at least see this information for myself haha.
PS - to make my point further, I hope people aren't reading the PSU section of this report and then assuming "Gee Fortron and Coolermaster must have the bestest PSUs evar!" when in reality all that is being presented are consumer return rates for this one retailer. The real thing to take away from this, maybe, are changes from year over year, but at the same time putting these numbers properly into perspective.
|
United Kingdom20344 Posts
Some numbers if anyone is interested:
+ Show Spoiler +Physics+Effects Maxed
I7 950 @3.8ghz -
Min, Max, Avg
19, 53, 28.933
3570k @5ghz -
Min Max Avg
32, 64, 45.133
Physics off, effects medium
i7 950 @3.8ghz -
Min, Max, Avg
27, 57, 40.600
3570k @5ghz -
Min, Max, Avg
44, 80, 60.333
|
Where are the frame latency numbers :p
These are 1080p and your usual variables.txt settings right?
|
United Kingdom20344 Posts
Where are the frame latency numbers
Can provide, they didn't seem neccesary though
1920x1080 and on med shaders, max textures, some other stuff low etc - but im pretty sure that's irrelevant, at least for the minimums
I'm so amazed by this chip going 5ghz on so low voltage, he said the hottest core throughout the entire day (absolute max temp) from gaming, LoL/sc2 load was 58c, like you could probably take this to 5.2ghz on just a hr-02 macho and a couple of case fans (if you dont mind temps poking at 90 absolute max's) or with a delid and a little higher end stuff, i wouldn't be suprised by 5.4ghz.. It's so crazy. Entire new batches of CPU's performing like this..
If Haswell has the IHS issues fixed and overclocks like this while having what appears to be ~8% IPC lead - plus the unlocked base clocks etc (still unclear if you can OC non-k-edition CPU's with it - but still*), just more settings, then it will be a massive upgrade from even sandy/ivy bridge, cmon intel don't fuck this up :D
*Secretly cheering for 5.4ghz i3's
|
On April 13 2013 07:49 skyR wrote:Show nested quote +On April 13 2013 01:34 Ata wrote:+ Show Spoiler +I was wondering how much you guys would put this on sale for:
560-Ti Twin Frozr
i5-2500K
Gigabyte Z68 D2H
Mushkin Enhanced Silverline Stiletto 8GB 2X4GB
XFX Core Edition Pro 450
Coolermaster HAF 912
Western Digital Caviar Blue 500GB
Samsung DVDRW 22X
Coolermaster Hyper 212+
Crucial M4 CT064M4SSD2 2.5" 64GB
Windows 7
In Canada, I quickly put togther a system from ncix with a 650 Ti Boost @ $165 (cheapest 7850 was $190 i think) and other similar components:
cheapest 3750k, $100 mobo, cheapest 64GB ssd, cheapest 500 GB harddrive, cheapest dvd... it came to $913 before tax and shipping and im thinking a -% of that.
I might be selling it to a coworker as im not really playing much.
Thanks Probably $700-$800
Thanks for the feedback
|
On April 14 2013 05:09 Gumbi wrote:
Mindfactory is really really good. Best prices in EU imo (pity outside from Germany it's generally no worth using cos case comes from separate warehouse so you have 30 euro shipping by 2). Boxed 3570k for €188 I've seen a few times on offer. Tat's absolutely insane for EU prices. Hardwareversand is very good too. Another German site, tends to be only a little bit more expensive, but a flat 19 euro shipping rate is very good.
When ordering from mindfactory, it makes no sense to order the case from them as well. I'd just order everything but the case and buy the case locally from somewhere. For example in Finland(which is like the most costly place in Europe for computer hardware), most cases are just going to be around 5-10 euros more so definitely worth it.
There are no early leaks / benchmarks of Haswell and or hard release dates yet?
|
^^Nope no legitimate leaks yet haha. And I would hardly call THG's preview as concrete either. Just gotta be patient for the NDA to finally pass in June.
|
On April 14 2013 08:36 mav451 wrote:Show nested quote +On April 14 2013 07:56 MrCon wrote:On April 14 2013 03:30 Ropid wrote:On April 14 2013 02:57 NihiLStarcraft wrote:On April 14 2013 02:38 Ropid wrote:
You can get a Samsung 840 Pro for 115 EUR or something like that. The Samsung 840 (without "Pro") is 145 EUR for 250 GB.
Comparing the OCZ Vertex4 with the Samsung 840 Pro, it looks like the Vertex4 is a bit faster both when it comes to reading and writing. And the 840 Pro is only 5-10 bucks cheaper! Not sure if it's worth it? That's not true. Anandtech does benchmarks where they try to simulate what happens in the real world, and the Samsung beats it handily there. I think the Samsung is actually the very best drive. I'm a bit suspicious of OCZ. They had a history of problems regarding reliability. They also have to buy the controller and memory they use, while Samsung can make everything themselves. That's where the better price can come from. It does not have to mean worse quality. Yeah, and even now OCZ has 5% defects versus 0.48% for Samsung for SSDs. Heh wasn't there some hubbub last year about people misinterpreting 'return' rates as failure rates http://www.hardware.fr/articles/881-7/ssd.htmlTom's, of all places, actually voiced the same concern I had about how to interpret this kind of data in their 2011 column. http://www.tomshardware.com/reviews/ssd-reliability-failure-rate,2923-3.html+ Show Spoiler +A drive failure implies the device is no longer functioning. However, returns can occur for a multitude of reasons. This presents a challenge because we don’t have any additional information on the returned drives—were they dead-on-arrival, did they stop working over time, or was there simply an incompatibility that prevented the customer from using the SSD?
This information only presents us with more questions. If online purchases account for the majority of hard drives sold, poor packaging and carrier mishandling can have a real effect on return rates. Furthermore, we also have no way of normalizing how customers used these drives. The large variance in hard drive return rates underlines this problem. For example, the Seagate Barracuda LP rises from 2.1% to 4.1%, while the Western Digital Caviar Green WD10EARS drops from 2.4% to 1.2%.
Alright, so the data really tells us nothing about reliability. What does it say then? Well, there seem to be more satisfied French customers purchasing Intel SSDs and not returning them than any other brand. Interestingly, in private chats, engineers from several SSD manufacturers note that a significant percentage of returned drives are unused and sealed, even though customers claim a compatibility issue. Customer satisfaction is interesting, but it's far less interesting to us than failure rates. Moving on. To be fair, for the more recent Oct 25, 2012 numbers, Hardware.fr does state that there is at least some kind of cross-checking (type de panne, tests croisés) to make that determination. It is rather odd that while they have access to this one French retailer's database they continue to make no effort at presenting us this data in its entirety. I would assume there's some understanding that they won't b/cc manufacturers are bound to get mad, but I would like to at least see this information for myself haha. PS - to make my point further, I hope people aren't reading the PSU section of this report and then assuming "Gee Fortron and Coolermaster must have the bestest PSUs evar!" when in reality all that is being presented are consumer return rates for this one retailer. The real thing to take away from this, maybe, are changes from year over year, but at the same time putting these numbers properly into perspective.
With all my respect to Toms, this is bullshit. They write :
A drive failure implies the device is no longer functioning. However, returns can occur for a multitude of reasons. This presents a challenge because we don’t have any additional information on the returned drives—were they dead-on-arrival, did they stop working over time, or was there simply an incompatibility that prevented the customer from using the SSD?
This information only presents us with more questions. If online purchases account for the majority of hard drives sold, poor packaging and carrier mishandling can have a real effect on return rates.
But this is unimportant, unless the number is so low that a single RMA skews the % a lot. And with numbers such a 0.48%, it doesn't seem so.
Why the product is RMAed is also unimportant, the simple question is : do you prefer a product that random users send back 5% of the time, or do you prefer the product they send back 0.5% of the time ? No more no less. It's not the first, or even 4th, criteria of choice, but when you hesitate between 2 products, it can make the difference. Especially when the numbers are around 5% or more for a long period (the OCZ was at 8% in the previous period of time), it usually show something is wrong, like with the IBM hard drives some years ago.
Hardware.fr is in the unique position of having access to these numbers, and having the power to publish them without fearing of retaliations (like a constructor removing his advertising). I'm pretty sure Toms&co wouldn't publish them because they depend too much on advertising.
|
Yeah I'm gonna have to disagree with you.
Have you actually dealt with dumb customers? There are far too many. As I said before, I would throw extreme caution to relying on these numbers and making broad generalizations with them. One retailer cannot hope to provide enough data points to make a good conclusion.
I would see it much differently if they had access to Europe, NA, SA, and Asia data as well. This is coming from a single French retailer. Do you not see the problem of such a narrow focus of customer return data? Realize I'm not saying to totally discount it, but it is irresponsible how many websites have quoted those numbers as gospel.
And just for clarity, I have stuck with WD Blues (or Blacks) and Samsung SSDs. But unless we know true failure rates and not simply 'return' rates, this is ONLY one more thing to consider. It is not THE thing. Please understand the distinction as it is critical in approaching this analysis correctly.
|
Sample size is too low for most categories to make strong claims about the nature of returns, much less failures. Something like 5% to 0.5% is probably statistically significant though.
|
Yeah - I can maybe get behind monitoring trends. But as I said earlier year over year changes could - the key word is could, hint at an increase/decrease in reliability. But these trends are purely return rates, not failure rates. And with little transparency in the data they are relaying to us, I cannot confidently put much weight or significance into the percentages.
|
|
|
Okay, so following up yesterdays awesome help and advice, I compiled the following shopping list today:
Minimum goal: Being able to stream SCII in 720p30fps fluently. Not looking to overclock.
Mobo ASRock B75M EUR 49 (@mindfactory.de)
CPU Intel Core i5-3570 EUR 198 (@mindfactory.de)
GPU Palit NVIDIA GTX 650 Ti EUR 121 (@amazon.de)
RAM TeamGroup Value 2x4GB DDR3 1600 EUR 44 (@mindfactory.de)
SSD Samsung 840 Basic 120GB SATA 3 EUR 79 (@mindfactory.de)
PSU XFX Core V2 450W EUR 42 (@mindfactory.de)
TOTAL EUR 533
As you can see, I found the cheapest price for every component on mindfactory.de except for the GPU, which I would get from amazon.de. So, before I order these parts (looking to buy in a couple of days), I wanted you experts to maybe take one last look at it and let me know if there's any way I could potentially save money etc. A couple questions: (1) I couldn't find much on the Palit GTX 650 Ti - it is cheaper even than the Zotac variety which has me feeling a tiny little bit uncomfortable. Anybody have any opinion on it? Is it a good idea to save 10 bucks here and not go for the Zotac? (2) Do I need a 450W PSU for my components or will less do? Looking to not pay more than necessary on my electric bills! 
Thanks for everything so far, this really is an invaluable resource!
|
Most of the cheapish Palit, Zotac, etc. graphics cards should be okay but nothing special at all. Aside from differences in warranty support, it's just in the cooler or maybe PCB design, no big deal for most people. Same GPU, so same performance. It's generally the Asus, MSI, maybe Gigabyte custom designs that tend to be better, though it varies by model.
You certainly don't need anything rated close to 450W max for that system. For many years, computer power supplies are switched-mode, with efficiency relatively constant across output loads (except under say 25% of max capacity or so). e.g. if computer needs 100W, power supply takes 118W from the wall (~85% efficiency); if computer needs 400W, power supply takes 482W from the wall (~83% efficiency) and so on. Depends on the design. But output power rating within reason, won't have much impact on power consumption. As for your system, you'd never use close to 200W; it would be closer to 100W in a game.
There isn't really anything cheaper than XFX Core V2 that is better, as reliable, so not much point in getting less. Maybe Enermax Triathlor 300W for 2 euros less? Warning: for some of the low-wattage options, they often give you short cables, which may not work with some cases.
|
On April 16 2013 02:30 Myrmidon wrote:
Most of the cheapish Palit, Zotac, etc. graphics cards should be okay but nothing special at all. Aside from differences in warranty support, it's just in the cooler or maybe PCB design, no big deal for most people. Same GPU, so same performance. It's generally the Asus, MSI, maybe Gigabyte custom designs that tend to be better, though it varies by model.
Okay, makes sense. In that case, I will go with the Palit and save the maximum amount of money!
On April 16 2013 02:30 Myrmidon wrote:
You certainly don't need anything rated close to 450W max for that system. For many years, computer power supplies are switched-mode, with efficiency relatively constant across output loads (except under say 25% of max capacity or so). e.g. if computer needs 100W, power supply takes 118W from the wall (~85% efficiency); if computer needs 400W, power supply takes 482W from the wall (~83% efficiency) and so on. Depends on the design. But output power rating within reason, won't have much impact on power consumption. As for your system, you'd never use close to 200W; it would be closer to 100W in a game.
There isn't really anything cheaper than XFX Core V2 that is better, as reliable, so not much point in getting less. Maybe Enermax Triathlor 300W for 2 euros less? Warning: for some of the low-wattage options, they often give you short cables, which may not work with some cases.
Ah, that makes perfect sense! I will stick with the XFX Core V2 for that price!
So that covers all my questions, let me know if anything in the build still looks bad/weak for the price or if you'd change anything - if nothing comes up, I will order in the next couple of days!
|
|
|
|
|
|
EPT
