OC3D Review i7 2600k, i5 2500k & 2300 1155 Sandy Bridge CPU's
- Thread starter tinytomlogan
- Start date
Filekutter
New member
TTL, I"d be interested to see if you could give a small idea over the time you use it, of any raise in your electric bill with that monster. hehe. Friend of mine has a 1000 Watt and I"ve heard its a socket hog from heck.
Mr Muggles
New member
I've asked Tom this on youtube actually ha ha His leccy bill must be massive.
People forget that when they buy these overkill 1000w PSU's that they are actually going to end up paying higher leccy bills as a result. Well over £200 more a year
killablade
New member
Correct, but a 1000w is more efficient stressed than when only 100-300w is being used. A 650w, for example, is usually more efficient at those loads. Therefore a 1000w psu would indeed use more power, simply because it's less efficient at those loads.
hmm how come you didn't do the obvious real life test on any sandy bridge ?
that being a x264 bench ?
if you compile x264 from git
a simple
make checkasm;./checkasm
checkasm –bench
would probably be a good start to 2011, you Probably want to disable SpeedStep though in the Bios to get consistent numbers OC.
if you get problems or want more help with making real life testing scripts etc then you can ask in http://doom10.org/index.php?topic=717.0 for instance or the IRC #x264 and #x264dev channel's where C/Assembly development happens (x86/64 and ARM etc) OC
that being a x264 bench ?
if you compile x264 from git
a simple
make checkasm;./checkasm
checkasm –bench
would probably be a good start to 2011, you Probably want to disable SpeedStep though in the Bios to get consistent numbers OC.
if you get problems or want more help with making real life testing scripts etc then you can ask in http://doom10.org/index.php?topic=717.0 for instance or the IRC #x264 and #x264dev channel's where C/Assembly development happens (x86/64 and ARM etc) OC
Technex
New member
Correct also
Most systems actually use only 200-300w or so even under load. Therefore a 600w PSU is most efficient in most cases.
I purchased my 700W as it was only a few quid more and gives a little more overhead
xenochrophis
New member
great reaview as always mate.
cheers.
cheers.
Tom ....love the video reviews... I'm learning so much!
The Sandybridge releases have thrown a spanner in my plans to build a new PC .. my first.
I'm not a gamer ... I need a multi-monitor (4x23") workstation for financial trading ... so only 2D graphics.
I was planning on using an i7-950 ... should I now consider the i5 2500K instead?
Many thanks in advance.
The Sandybridge releases have thrown a spanner in my plans to build a new PC .. my first.
I'm not a gamer ... I need a multi-monitor (4x23") workstation for financial trading ... so only 2D graphics.
I was planning on using an i7-950 ... should I now consider the i5 2500K instead?
Many thanks in advance.
No mate it wont make any difference to you what ever one tbh dude. Your work isnt going to be massively intensive and you wont be gaming / rendering.
Tom Logan, I have recently purchased an I7 970 from eBay for 500$ a week ago NIB. Did I make a bad decision by not waiting for the 2600k? I have the Rampage 3 Extreme + corsair dominator gt and I love my setup at the moment but after watching and reading your review I feel I should of waited. Any thoughts ?
P.S. http://www.cpubenchmark.net/high_end_cpus.html Accord to this website the 2600k for 330$ almost out preforms the i7 970. Is that website correct or did they ramp up the multiplier to get those results.
P.S. http://www.cpubenchmark.net/high_end_cpus.html Accord to this website the 2600k for 330$ almost out preforms the i7 970. Is that website correct or did they ramp up the multiplier to get those results.
Maybe I'm just not getting it. I don't understand what all the fuss is about with the 1155's. I saw in one review that they don't have hyperthreading then read in another that they do, theydon't support more than 4x AA, they don't support above DX 10.1, they don't support cuda or physx, the Intel drivers aren't any good,are still immature. What good is anintegrated graphics core if it has all these short comings? Why do you even need an integrated graphics core, isn't that what the GPU does?? The IGC automatically gets disabled when you install a discreet graphics card (so your paying for substandard technology that wont ever be beneficial to most of us anyway??)
If I'm misinformed or you can shed some light on all the overhype, I'd appreciate it.
EDIT...Just saw that the I5's don't have hyperthreading and the I7's do have it, but other problems still exist
If I'm misinformed or you can shed some light on all the overhype, I'd appreciate it.
EDIT...Just saw that the I5's don't have hyperthreading and the I7's do have it, but other problems still exist
I think your missing the big picture here caveman. MOST people here aren't buying the cpu for the igc. They are buying it because at the pricepoint they come in at, for example in the U.S. the 2500k comes in at $220. Ok, at that price range you have two options, AMD 1090T x6 or or i5-760 (which doesn't have HT anyway), or for a few bucks more the i7-870. Well the 2500k smokes these as well as the i7-950 ($299) at stock settings, even without having HT. The 2500k is perfect if you are building a new rig. I don't think anyone would recommend switching over if you already have an i7 1366 setup. Now coupled with the fact, you can OC the hell out of the turbo mode on these sandy bridge chips, for a true quadcore to get over 4.5ghz is friggin awesome for $220. Perfect for mainstream gamers, who lets face it, will have a discreet graphics card anyway...or those who want to upgrade, or building their first system. It is my understanding that most games will take advantage of only 4 cores as of right now, and that a true quadcore will perform better than a 4c/8t cpu when only 4 threads are being used. Now as far as the htpc crowd, the lower end variants of the SB chips will be more than adequate for their home theater needs, without having to spring for a low end discreet graphics solution. So if you're viewing these sandy bridge chips from the top, i7 950/970/980x....then yes, theres really nothing in them for you. But if you're starting a fresh build, you'd be a fool to overlook them.
Iirc the 2600 has HT.
Be aware that putting anything more than 1.4v through the chip will most likely kill it as the 32nm chips simply cannot handle that voltage and the IMC will fry. Numerous reports are beginning to surface of sandys dying because people are treating them like their 45nm counterparts. The new chips do look very nice though but sadly, as stated above, there is little point in switching (note I didn't say upgrading) from i7 to Sandybridge because clock for clock, the performance is still with bloomfield/gulftown. The fact that these can clock to 5ghz is great, however I dare say if you put silly amounts of voltage through most cpu's, they will clock well. How long they will last though is an entirely different question. Nobody knows for sure at present what voltage is deemed safe. Intel say 1.35, overclocking forums suggest 1.4v as an absolute max. I have just read about one of these cpu's dying with 1.4 so in reality, nobody can say for sure!
Also getting 5ghz from a cpu is great (congrats Tom). However, Asus claim (so I have read) that only 10% of the cpu's they have had reached this. 30-40% would do 4.5ghz with the remaining 50 being 4-4.5ghz. All this is presumebly with excessive voltage which as stated above, 32nm archs don't like at all and will die as a result. I dare say the chips you have tom are on their last legs already, maybe dead?
Sorry gents, but I aint getting all excited over a cpu that:
1. Is not as fast as 1366 i7.
2. Suceptable to failure with mild (0.2) voltage increases
3. No HT on the entry level models
Great for someone looking for a new build, upgrade from 1156 perhaps but a sidestep for 3 year old bloomfield. I really hope bulldozer offers more than this. Disappointing.
Be aware that putting anything more than 1.4v through the chip will most likely kill it as the 32nm chips simply cannot handle that voltage and the IMC will fry. Numerous reports are beginning to surface of sandys dying because people are treating them like their 45nm counterparts. The new chips do look very nice though but sadly, as stated above, there is little point in switching (note I didn't say upgrading) from i7 to Sandybridge because clock for clock, the performance is still with bloomfield/gulftown. The fact that these can clock to 5ghz is great, however I dare say if you put silly amounts of voltage through most cpu's, they will clock well. How long they will last though is an entirely different question. Nobody knows for sure at present what voltage is deemed safe. Intel say 1.35, overclocking forums suggest 1.4v as an absolute max. I have just read about one of these cpu's dying with 1.4 so in reality, nobody can say for sure!
Also getting 5ghz from a cpu is great (congrats Tom). However, Asus claim (so I have read) that only 10% of the cpu's they have had reached this. 30-40% would do 4.5ghz with the remaining 50 being 4-4.5ghz. All this is presumebly with excessive voltage which as stated above, 32nm archs don't like at all and will die as a result. I dare say the chips you have tom are on their last legs already, maybe dead?
Sorry gents, but I aint getting all excited over a cpu that:
1. Is not as fast as 1366 i7.
2. Suceptable to failure with mild (0.2) voltage increases
3. No HT on the entry level models
Great for someone looking for a new build, upgrade from 1156 perhaps but a sidestep for 3 year old bloomfield. I really hope bulldozer offers more than this. Disappointing.
OC when you realise that all AMD to date run assembly slower than Intel then id like to see IF bulldozer etc can finally improve on that in the future.
that's the main reason Why im asking for this simple x264 C/assembly code test
make checkasm;./checkasm
checkasm –bench
its the only real life benchmark that will show You if things have changed, plus OC i cant seem to find any SB standard clocked and over clocked pastebin results to date to compare to older CPU's never mind setting a baseline for future products and that's a shame.
Yeah I too would prefer to see benchmarks clock for clock, not just stock and overclocked. This would enable the end user at least to differentiate between which cpu performs best albeit on a clocking level. Thise of you who remember the slower clocked Opteron/x2 prcoessors vs higher clocked Pentiums will know that the Opty was superior in most ways. Clockspeed isn't everything but when we are talking similar archetectures then I think it's a viable comparison.
Similar threads
