Phenom Benchmark at Ocworkbench

Is this the real deal or the Phenom could get better?

http://www.ocworkbench.com/2007/gigabyte/GA-MA790FX-DQ6/b1.htm
56 answers Last reply
More about phenom benchmark ocworkbench
  1. keep it real said:
    Is this the real deal or the Phenom could get better?

    http://www.ocworkbench.com/2007/gigabyte/GA-MA790FX-DQ6/b1.htm



    It looks like Intel has an adantage only in Multimedia. The funny thing is that Anand showed that K10 is better in SSE that isn't Intel optimized.
  2. BaronMatrix said:
    It looks like Intel has an adantage only in Multimedia. The funny thing is that Anand showed that K10 is better in SSE that isn't Intel optimized.


    I´m gonna wait for some more benches too...

    Anyway there´s something I already know: I´m getting Phenom (I like AMD and I don´t overclock)
  3. These tests were preformed at the same clock speed (they underclocked the Intel chip to get it down to the AMD's speed)

    Unless I am misunderstanding the tests...

    The Intel had a 23% advantage in Drystone
    The AMD Had a 2% advantage in Whetstone
    The Intel had a 188% (!) advantage in Integer Multimedia (ie games)
    The Intel had a 20% advantage in Floating point multimedia

    In the gaming tests. The Intel won every test except one....

    The best overclock the AMD chip would give was 2.53 ghz...

    Unlike what Baron has posted, it seems Intel has an advantage in every test except Floating point multi-media where things are basically tied.

    Ouch - AMD is in deep sh*t if these numbers are true.

    I would also note this is a test that compares a new, barely released vapourware AMD chip to the old, being phased out Intel quad.

    The new penryn core is 7% faster clock for clock (or so) and the 45 nano Quads are already doing 4.0 ghz + on air cooling all over the place.

    If these test are accurate a 3.0 gh Penny quad shuld win every benchmark in a walk.

    This could get really, really ugly for AMD.

    Can you imagine what Nehalem will do?.. tick... tock... tick... tock...


    (ok... I will stop taunting Baron now........)
  4. It looks like Intel won or held even with a year-old chip in whatever was tested, and on top of that, don't forget clock speeds. The test involved downclocking the FSB on the QX6700 to normalize the CPU frequencies. Even their cheapest quad (Q6600) comes stock with a higher frequency and bus speed.

    I don't understand how the author concluded with this:
    Quote:
    In fact, If you look at the benchmarks of games, the numbers aren't that far off from the Intel platform. The only more noticeable difference is in the multimedia encoding. In fact, the performance levels of the phenom was better than I expected.

    To me, performance was even worse than expected. Phenom didn't win any of the SSE4-enhanced benchmarks that I expected it to, yet QX6700 doesn't even support SSE4 like the Penryns just coming out. And we've seen the drastic effect the Penryns have on SSE4 benches.

    The similarity of the game benchmarks is simply due to GPU limitations. Whatever CPU weaknesses might be exposed in future gaming was masked by the use of high resolution (1600x1200) resulting in low frame rate (10 - 60 fps) on current cards. Even 3dmark06 revealed the Phenom setup was slower because of CPU, not so much graphics/MB.

    From someone on their forums expressing my sentiment:
    Quote:
    So it's:
    Slower clock-for-clock than Kentsfield, let alone Yorkfield.
    Slower clocked than Q6600.
    Doesn't overclock at all well. Kentsfield overclocks much better, Yorkfield even more so.

    It's going to be fighting a bit of an uphill struggle against Intel, isn't it? It'll have to be cheap...
  5. The QX6700 is running at a lower than stock FSB as well, 230MHz instead of 266MHz, so it is disadvantaged somewhat.

    This clock for clock testing isn't very useful except to prove that Phenom is behind in IPC to a FSB limited C2Q. When we have 45nm C2Qs overclocking to 4GHz, and all they eek out of the Phenom is 2.53GHz?!

    Also, I'm not sure what numbers Baron is looking at, but C2Q is clearly faster clock for clock.
  6. Quote:
    In our test, we have to use Gigabyte's own single slot 2600XT
    That explains why the game scores were so close together, GPU limited all the way...
  7. Wonder how they pumped out the 4.13 and 4.35 scaling in Cinebench
  8. poor poor AMD, there 3870 cant keep up with the 8800gt and their phenom cant match the C2D. I guess its time for me to move over to Intel again. So funny all this. cos i currently run AMD+ Nvidia, when i switch its gonna be Intel+ATi haha
  9. cruiseoveride said:
    poor poor AMD, there 3870 cant keep up with the 8800gt and their phenom cant match the C2D. I guess its time for me to move over to Intel again. So funny all this. cos i currently run AMD+ Nvidia, when i switch its gonna be Intel+ATi haha


    As a side note I love the 3870 series. Dirt cheap and it performs GREAT at it's price point.
  10. Kari said:
    Quote:
    In our test, we have to use Gigabyte's own single slot 2600XT
    That explains why the game scores were so close together, GPU limited all the way...


    Definitely agreed. The games were tested at 1600 x 1200.... I even wonder if OCWorkbench knows how to test CPU with games...

    Other discrepancies: Multimedia Int x8 test for Phenom should be discarded, since if I remembered correctly, Phenom can only support up to Int x4. Super Pi 1M should also be disregarded, as Super Pi was not written to take advantage of AMD's SSE engine. AMD has always performed a lot slower in Super Pi since K8 than Intel's Core 2, clock for clock.
  11. the_vorlon said:
    These tests were preformed at the same clock speed (they underclocked the Intel chip to get it down to the AMD's speed)

    Unless I am misunderstanding the tests...

    The Intel had a 23% advantage in Drystone
    The AMD Had a 2% advantage in Whetstone
    The Intel had a 188% (!) advantage in Integer Multimedia (ie games)
    The Intel had a 20% advantage in Floating point multimedia

    In the gaming tests. The Intel won every test except one....

    The best overclock the AMD chip would give was 2.53 ghz...

    Unlike what Baron has posted, it seems Intel has an advantage in every test except Floating point multi-media where things are basically tied.

    Ouch - AMD is in deep sh*t if these numbers are true.

    I would also note this is a test that compares a new, barely released vapourware AMD chip to the old, being phased out Intel quad.

    The new penryn core is 7% faster clock for clock (or so) and the 45 nano Quads are already doing 4.0 ghz + on air cooling all over the place.

    If these test are accurate a 3.0 gh Penny quad shuld win every benchmark in a walk.

    This could get really, really ugly for AMD.

    Can you imagine what Nehalem will do?.. tick... tock... tick... tock...


    (ok... I will stop taunting Baron now........)



    When I play games, I'm not concerned about whose fastest. For me a 2.2GHz quad will be probably 70% faster than my X2 4400+. It's just that the vitriol displayed now that AMD is SLIGHTLY behind is sickening. We all should have head Intel's head in a bag in 2005, but we let it go and piled up with NetBurst. Even the OEMs didn't try to advance their AMD presence until last year - after Intel released a better chip.

    I'm not upset about it because Intel should win with 4 issue. I don't think it's cheating but it is something to consider. I would think that BullDozer will be more than 3 issue to override that advantage. Also, in all the benches I've seen, they pretty much all have different numbers and AMD always increases perf over subsequent revisions.

    For the first iteration, they're doing well.
  12. keep it real said:
    Is this the real deal or the Phenom could get better?

    http://www.ocworkbench.com/2007/gigabyte/GA-MA790FX-DQ6/b1.htm

    Pretty good and it looks like Intel has no worries as to who is the king of the mountain. AMD has gained some ground tho on Intel as now it will take the quads to beat out Phenom. Before the best AMD could do is match the performance of the E6550 with the X2 6400+.

    I'm not sure how the 4X 3870 GPU's on an AM2+ with a phenom 2.4GHz will play against Intel but gaming may not be an easy win for Intel. These options are in a way a counter to Intel's CPU being the king of the hill. We are yet to see any Phenom FX which may be 2.6GHz and on a 4X4 could make for some interesting twists. I know the 4X4 isn't a major threat to Intel as it has its draw backs but do we really know what kind of performance to expect for FX?

    No matter the case I think AMD cant compete against Intel's massively high OC's presently. The lack of any drivers to push the 4X 9600 could make this a small stick for AMD at best. The biggest threat AMD may hold is a possible 2X X4 Phenom kentsfield style come around May. I don't think Intel can do a 4 way package of their Penryn dual core. AMD taking advantage of 8 cores on one socket may rest in their ability to move to 45nm which due to past delay performance may end up being against a 8 core Nehalem.
  13. CPU market will end up just like the GPU Market. Replace nVidia with Intel.

    DAAMIT will offer great price/performance for mainstream people and will do just fine.

    Intel will offer great MAX performance and price gouge for it.

    Yes the current Penryn at $1200 is going 4 GHz. However the $300-400 Penryn will have an 8x Multi. Good luck getting 500 FSB out of the quad. The $500-600 will have an 8.5. It's also rediculous to charge an extra $200-300 bucks for an extra half step.

    That said, I'll be getting a Q9450 for SSE4 support.. Not this SSE4.0a stuff.

    I will however be getting ATI cards. I don't want to live in a world with just Intel/nVidia... so I will support DAAMIT in the market that I see most fit. I hate nVidia.
  14. keep it real said:
    Is this the real deal or the Phenom could get better?

    http://www.ocworkbench.com/2007/gigabyte/GA-MA790FX-DQ6/b1.htm

    AMDs performance isn't as bad as i expected. The prices they ask for it are though. While the mainboards for AMD tend to be a little cheaper compared to adequate intel mainboards (thanks to the IMC) the memory on the AMD platform needs to be better, thus more expensive.
    The performance offered for multi-socket platforms is quite good. Sadly, the prices for the single socket platforms aren't nearly competitive with intels offerings. And that is without considering overclocking. AMD should look at their GPU department or their dual-core chips to get an idea on how to price those Phenoms.
  15. cnumartyr said:
    CPU market will end up just like the GPU Market. Replace nVidia with Intel.

    DAAMIT will offer great price/performance for mainstream people and will do just fine.

    Intel will offer great MAX performance and price gouge for it.

    Yes the current Penryn at $1200 is going 4 GHz. However the $300-400 Penryn will have an 8x Multi. Good luck getting 500 FSB out of the quad. The $500-600 will have an 8.5. It's also rediculous to charge an extra $200-300 bucks for an extra half step.

    That said, I'll be getting a Q9450 for SSE4 support.. Not this SSE4.0a stuff.

    I will however be getting ATI cards. I don't want to live in a world with just Intel/nVidia... so I will support DAAMIT in the market that I see most fit. I hate nVidia.


    Many P35 boards can get over 450 FSB. X38/X48 probably has a decent shot at 500FSB with a mature BIOS, even right now they are doing around 485FSB.

    The Phenoms still seems overpriced relative to the Q6600 and especially compared to the upcoming Q9300/9450.
  16. better be better
  17. It looked to me like AMD will be competitive in the non-OC market. The game benches were close, but AMD definitely has to up that clock speed.
  18. San Pedro said:
    It looked to me like AMD will be competitive in the non-OC market. The game benches were close, but AMD definitely has to up that clock speed.


    Of course it would be close in gaming, even a Sempron or Celeron would be competitive with a 2600XT acting as a bottleneck. :lol:
  19. BaronMatrix said:
    When I play games, I'm not concerned about whose fastest. For me a 2.2GHz quad will be probably 70% faster than my X2 4400+.

    That is, if the game supports multi-threading up to four cores. I'm not sure if Phenom 2.2Ghz will be faster than your X2 4400 in Crysis.

    Quote:
    It's just that the vitriol displayed now that AMD is SLIGHTLY behind is sickening. We all should have head Intel's head in a bag in 2005, but we let it go and piled up with NetBurst. Even the OEMs didn't try to advance their AMD presence until last year - after Intel released a better chip.

    What you're seeing here is when a company promised something, and didn't deliver. Although I don't want to hammer this, but where is the 40% better in performance? Where is the supposedly 2.2Ghz Phenom performing on par with Q6600?

    The OEMs know AMD doesn't really have the capacity to supply them with sufficient amount of chips. Take a look at Barcelona now: where are they?

    Quote:
    I'm not upset about it because Intel should win with 4 issue. I don't think it's cheating but it is something to consider. I would think that BullDozer will be more than 3 issue to override that advantage. Also, in all the benches I've seen, they pretty much all have different numbers and AMD always increases perf over subsequent revisions.

    So let me get this straight... you think a 4 issue execution engine is considered cheating?
    Maybe I should start considering AMD having IMC as cheating too.

    Quote:

    For the first iteration, they're doing well.

    For the first iteration, they're not living up to the hype. After delaying the product for half year, and it still cannot match up with a product that's one year old. Intel has lived up to its hype, but has AMD?
  20. BaronMatrix said:
    It looks like Intel has an adantage only in Multimedia. The funny thing is that Anand showed that K10 is better in SSE that isn't Intel optimized.



    Nice try fanboy. The Intel proc was underclocked to the speed of the phenom.
  21. I'm confused. Why is the Memory bandwith so low with the 9600? My 6000+ with ddr2 800 hits 9700+
    in Sandra. I thought the new phenoms had a new memory controller and with ddr2 1066 the bandwith should be double that?
  22. BaronMatrix said:
    When I play games, I'm not concerned about whose fastest. For me a 2.2GHz quad will be probably 70% faster than my X2 4400+. It's just that the vitriol displayed now that AMD is SLIGHTLY behind is sickening. We all should have head Intel's head in a bag in 2005, but we let it go and piled up with NetBurst. Even the OEMs didn't try to advance their AMD presence until last year - after Intel released a better chip.

    I'm not upset about it because Intel should win with 4 issue. I don't think it's cheating but it is something to consider. I would think that BullDozer will be more than 3 issue to override that advantage. Also, in all the benches I've seen, they pretty much all have different numbers and AMD always increases perf over subsequent revisions.

    For the first iteration, they're doing well.


    There is no vitriol..

    But as a consumer.. I do take the ultimate revenge I can in a free and open marketplace... I vote with my wallet.

    When the Wilammette core (P4 in the 1.3 => 2.0 ghz range) I bought T-Birds because they were a better deal (Faster, cheap DDR versus expensive Rambus)

    When the Northwood core (P4's 2.0 => 3.46 ghz) scaled like crazy and went to DDR, I bought northwoods.

    When Prescott came out, I stopped buying P4s and bought Athlons.

    When the Core2 got sane in pricing.. I went to the Core2s.

    I buy the best deal I can buy TODAY.

    And today that's a Core2

    Tomorrow that may change, and if it does, so will I.
  23. yomamafor1 said:
    That is, if the game supports multi-threading up to four cores. I'm not sure if Phenom 2.2Ghz will be faster than your X2 4400 in Crysis.

    Quote:
    It's just that the vitriol displayed now that AMD is SLIGHTLY behind is sickening. We all should have head Intel's head in a bag in 2005, but we let it go and piled up with NetBurst. Even the OEMs didn't try to advance their AMD presence until last year - after Intel released a better chip.

    What you're seeing here is when a company promised something, and didn't deliver. Although I don't want to hammer this, but where is the 40% better in performance? Where is the supposedly 2.2Ghz Phenom performing on par with Q6600?

    The OEMs know AMD doesn't really have the capacity to supply them with sufficient amount of chips. Take a look at Barcelona now: where are they?

    Quote:
    I'm not upset about it because Intel should win with 4 issue. I don't think it's cheating but it is something to consider. I would think that BullDozer will be more than 3 issue to override that advantage. Also, in all the benches I've seen, they pretty much all have different numbers and AMD always increases perf over subsequent revisions.

    So let me get this straight... you think a 4 issue execution engine is considered cheating?
    Maybe I should start considering AMD having IMC as cheating too.


    For the first iteration, they're not living up to the hype. After delaying the product for half year, and it still cannot match up with a product that's one year old. Intel has lived up to its hype, but has AMD?



    If you look at the tests Anand did you will see that AMD delivered EXACTLY what they said. "Outperform Intel by 40% on a variety of workloads." They didn't say EVERY test. AMD still rules HPC and apps that use heavy memory accesses. I didn't say it was cheating to have 4 issue.

    You can't compare products that way since it takes a year or two to design and qualify a CPU. It's not like they could have stopped midstream. I still say that people are holding AMD to a higher standard in some weird way. There are two x86 manufs. One is very large the other is not. Can we at least agree that Intel has improved Core 2 and that the delay meant Intel had extra time, just as AMD had with Opteron and subsequent revisions?

    I won't say that DDR2-1066 and retail boards and BIOS' will improve the scores tremendously, but I have yet to see a full batch of tests with actual retail HW running at spec. But remember, I am not concerned about one company being much faster than another, but about both being respected.
  24. TechnologyCoordinator said:
    Nice try fanboy. The Intel proc was underclocked to the speed of the phenom.



    I'm not a fanboy. I hate everyone equally.
  25. no comment
  26. BaronMatrix said:
    I'm not a fanboy. I hate everyone equally.


    But you care to admit the you tried to tout phenom as superior but the Intel proc was severely underclocked in the benchmark, not fair, huh?
  27. Results are ok, as in AMD is very close to C2Q in performance clock for clock. Penryn improvements isnt the biggest AMD headache atm, its clockspeed. Phenom FX 2.6GHz vs QX9775 3.2GHz is no match.

    If AMD could at least have high yields and flood the market as with ATI 3800 series, at least that would bring some profits but it doesnt seem to be the case.
  28. the_vorlon said:

    The Intel had a 188% (!) advantage in Integer Multimedia (ie games)


    Don't make me laugh...you are ignorant...did you not look at one of those gaming benchmarks??? They were all neck for neck on the fps; with AMD scoring 1-2 fps higher once or twice. So that means that there is no practical difference between the two in gaming clock for clock. However, if you OC the two chips, intel will get higher. But we do know that the AMD chips will get to 3.0Ghz because of earlier benchmarks of the chips from other websites. So the OC issue is not the processor, but rather the motherboard that is having issues. Hopefully AMD will get that fixed.
  29. BaronMatrix said:
    I'm not a fanboy. I hate everyone equally.

    You sure about that? :lol: :lol: :lol:

  30. TechnologyCoordinator said:
    But you care to admit the you tried to tout phenom as superior but the Intel proc was severely underclocked in the benchmark, not fair, huh?


    If I remembered correctly, he once stated that 1Ghz Barcelona would trounce Intel's any offering...

    Of course, I have a life, and don't really have time to go through that nasty history... :lol: :lol:
  31. Oi, no stealing my pics! There's 5 minutes of hard work there! :fou: :lol:

    Btw, unless I'm blind, all of the systems performed within 2-3fps of each other in games anyway, so if you clocked the QX6700 back to 2.66GHz, you'd probably widen the gap to 3-5fps. Gee, the Phenom got hammered :sarcastic:

    Seriously, if you don't intend to overclock, the Phenom is still a decent option.
  32. Harrisson said:
    Results are ok, as in AMD is very close to C2Q in performance clock for clock.


    Clock for clock doesn't matter one bit. What matters is performance. People aren't going to underclock their brand new Intel processors out of modesty.


    Clock for clock comparisons are only good for comparing architectures. When you want to compare performance, you run them at speed.
  33. Quote:
    Don't make me laugh...you are ignorant...did you not look at one of those gaming benchmarks??? They were all neck for neck on the fps; with AMD scoring 1-2 fps higher once or twice. So that means that there is no practical difference between the two in gaming clock for clock. However, if you OC the two chips, intel will get higher.

    The bolded part sounds incorrect. The games were benched at 1600x1200, yet the graphics card was a sluggish ATI 2600XT. It doesn't matter how well each chip overclocks; scores wouldn't have changed with that kind of budget card.

    That does not mean the chips are identical in gaming. Many people use the much faster 8800gt/gts/gtx, or multiple cards, and still others like to keep their CPU for the next generation of video cards, so gaming performance at artificially lowered resolutions is still useful information. This concern was not directly addressed.

    Quote:
    Baronmatrix:

    If you look at the tests Anand did you will see that AMD delivered EXACTLY what they said. "Outperform Intel by 40% on a variety of workloads." They didn't say EVERY test. AMD still rules HPC and apps that use heavy memory accesses.

    I don't see such a conclusion from Anand. They compared K10 to K8 - identical clocks, identical # of cores - and found improvements of up to ~20% but averaging 15% on desktop-centric benches. http://anandtech.com/cpuchipsets/showdoc.aspx?i=3092&p=5

    They then said this would not quite be enough to keep up with 65-nm Core 2, let alone clock speed differences and the improvements with 45-nm.
  34. BaronMatrix said:
    If you look at the tests Anand did you will see that AMD delivered EXACTLY what they said. "Outperform Intel by 40% on a variety of workloads." They didn't say EVERY test. AMD still rules HPC and apps that use heavy memory accesses.

    Please show me, in which of the test, Barcelona actually has 40% over Clovertown, clock for clock?
    For your convenience, I've also included the url of the review.
    http://www.anandtech.com/IT/showdoc.aspx?i=3091&p=1

    And in order to refresh your memory, let me also include a report on Randy Allen's statement of "40% better".
    http://www.news.com/AMD-Go-to-Barcelona-over-Clovertown/2100-1006_3-6152645.html

    Quote:

    But AMD's "Barcelona" quad-core chip, due to arrive midway through 2007, will be a significant notch faster than the Clovertown chips expected to be on the market at that time, said Randy Allen, AMD's corporate vice president for server and workstation products.

    "We expect across a wide variety of workloads for Barcelona to outperform Clovertown by 40 percent," Allen said. The quad-core chip also will outperform AMD's current dual-core Opterons on "floating point" mathematical calculations by a factor of 3.6 at the same clock rate, he said.


    I'm waiting, Mr. Christian Howell.

    Quote:

    I didn't say it was cheating to have 4 issue.

    I'm sorry. I think you meant, "although having a 4 issue execution core is not "cheating", per se, but I think you should take it into consideration, as it may make the comparison unfair".

    Quote:
    You can't compare products that way since it takes a year or two to design and qualify a CPU. It's not like they could have stopped midstream. I still say that people are holding AMD to a higher standard in some weird way. There are two x86 manufs. One is very large the other is not. Can we at least agree that Intel has improved Core 2 and that the delay meant Intel had extra time, just as AMD had with Opteron and subsequent revisions?

    Are you serious in that statement?
    Let's see, when Barcelona didn't deliver, did reviewers reflect that? No. When Barcelona didn't have the supposed availability, did reviewers jump on that? No.

    Do you think if the same thing happened to Intel, the same group of people will still hold that standard? No. Intel would be roasted, toasted, and grilled. Go and read some of the reviews from Intel's Crapburst days. But why didn't this happen to AMD?

    I agree that Intel had improved Core 2, and AMD's delay means more time for Intel to tweak Core 2. But is this Intel's fault?

    Quote:

    I won't say that DDR2-1066 and retail boards and BIOS' will improve the scores tremendously, but I have yet to see a full batch of tests with actual retail HW running at spec. But remember, I am not concerned about one company being much faster than another, but about both being respected.

    I would say DDR2-1066 will improve Phenom's performance, but I'm not sure how much it will influence. In game, it translated to about 8FPS, which closes the gap between Phenom and Core 2.

    We've already starting to see leaked reviews to trickle out, and its pretty consistent with what we've seen 2 months ago. Phenom will be competitive, price wise, if not overclocked. But with a little tweaking, Penryn will not only run faster, but do so with less power, and less heat.
  35. randomizer said:
    Oi, no stealing my pics! There's 5 minutes of hard work there! :fou: :lol:

    Btw, unless I'm blind, all of the systems performed within 2-3fps of each other in games anyway, so if you clocked the QX6700 back to 2.66GHz, you'd probably widen the gap to 3-5fps. Gee, the Phenom got hammered :sarcastic:

    Seriously, if you don't intend to overclock, the Phenom is still a decent option.


    Actually, as wr pointed out above, the resolution was set at 16 x 12. So it would be more GPU limited than CPU limited.
  36. Quote:

    Ouch - AMD is in deep sh*t if these numbers are true.


    Nope. AMD would be in deep s*it if:
    1. Yields are low enough such that the CPU + motherboard combination sells for more than an Intel set of comparable performance.
    2. They cannot make enough CPUs to meet demand.
    3. They are forced to sell the CPUs at a low enough price such that they don't make any money.
    4. The CPUs run significantly hotter than Intel's or they draw significantly more power.

    The performance is rather in-line with the competition in the actual (non-synthetic) tests for 95%+ of the market. As long as AMD can make them at a reasonable enough price so that they can make money AND they can make enough of them, they'll do just fine. Plus, desktops are the #3 priority at AMD. The server comes first as that is where AMD's margins and their performance lead is the greatest. Mobile is #2 as it is the largest and fastest-growing market in the U.S. This is very apparent as the 10h architecture was made for the server first and foremost. AMD is making a separate branch of development for mobile: the Griffin CPU, which is part K8, part 10h, and part new design. The desktop chip is basically a single-socket version of the Opteron.

    Quote:
    I would also note this is a test that compares a new, barely released vapourware AMD chip to the old, being phased out Intel quad.


    Apparently the new version isn't too much different from the old one. I think I saw that somewhere...

    Quote:
    The new penryn core is 7% faster clock for clock (or so)


    Ah, there it was.

    Quote:
    and the 45 nano Quads are already doing 4.0 ghz + on air cooling all over the place.


    That doesn't matter for almost all of the market as most desktop CPUs are in non-overclockable OEM units. The only part of the OC results that is meaningful is that Intel might have some extra headroom, which means that midrange and lower-end CPUs can get faster at similar price levels.

    Quote:
    Can you imagine what Nehalem will do?.. tick... tock... tick... tock...


    I don't know if anybody really does. It is a big change for Intel, not only the microarchitecture but also the entire platform and the CPU's macroarchitecture. AMD has already migrated from an FSB to an IMC and has over 5 years of experience in doing so. They have also been going the big-monolithic-die route, which Intel will supposedly behave to do as well. It will be interesting to see how well Intel executes.


    (ok... I will stop taunting Baron now........)
  37. yomamafor1 said:
    Actually, as wr pointed out above, the resolution was set at 16 x 12. So it would be more GPU limited than CPU limited.

    Most games are GPU limited. I can't see how running a QX6700 or a Phenom is going to make much difference in Crysis. I'd take a boxed Phenom 9500 at $300 over an OEM $700-1000+ QX6700 any day.
  38. randomizer said:
    Most games are GPU limited. I can't see how running a QX6700 or a Phenom is going to make much difference in Crysis. I'd take a boxed Phenom 9500 at $300 over an OEM $700-1000+ QX6700 any day.


    But I'll take the 300 USD for Phenom, buy a Q6600, and get a soda with the spare change :lol:
  39. I'll buy an E4400, and get 20 sodas with the change :D
  40. randomizer said:
    I'll buy an E4400, and get 20 sodas with the change :D


    Yes, but do you have enough cache space for Minesweeper? :pt1cable:

    I wonder how fast Minesweeper will run if I put the entire program inside L2 :lol:
  41. You can't fit it in L1?
  42. randomizer said:
    I'll buy an E4400, and get 20 sodas with the change :D


    If I were you, I'd skip the soda's and buy a life instead. :kaola: Just kidding. I'd have to say, if I were putting a quad core in my computer, I would be playing games at 16x12 or higher, so a quad AMD compared to a quad Intel wouldn't make a huge difference in that department. Intel has the better CPU architecture right now IMO, but I don't think AMD will go through the tubes anytime soon; mainly because us AMD fanboy's are so darn stubborn we'll buy the lesser of two products just to support or favorite company.
  43. I buy whatever I can afford at the time with the best performance at the time, ie. currently nothing, both for CPU and GPU.
  44. I look at the whole debate like this:

    I have this much to spend on a complete system. Who will give me the best performance for my money? I really don't give a care if it's AMD or Intel, I just want the best performance I can afford.
  45. Exactly.
  46. INdeed. However right now I dont like how 8800GT's are going faulty and have heat issues. Not to mention they are overpriced and not in stock. And I dont like any chipsets for intel either. I dont want DDR3...And I want to SLI or Crossfire.

    I want best price/perfomance and right now Its looking to be the phenom X4 9700, a 790fx chipset, and 2 HD3870's(maybe an extra HD3850 thrown in too :D )

    Unless waiting another 2 months for a penryn to be priced around 300bucks to come around is worth it....And to wait for the G92 8800GTS....or wait for the HD3870 x2.

    I feel like my head is going to explode :ouch:
  47. BaronMatrix said:
    AMD still rules HPC and apps that use heavy memory accesses.

    Like this? http://www.spec.org/cpu2006/results/res2007q3/cpu2006-20070903-01948.html
  48. Oh, my eyes are non-compliant!
  49. randomizer said:
    Oi, no stealing my pics! There's 5 minutes of hard work there! :fou: :lol:

    Btw, unless I'm blind, all of the systems performed within 2-3fps of each other in games anyway, so if you clocked the QX6700 back to 2.66GHz, you'd probably widen the gap to 3-5fps. Gee, the Phenom got hammered :sarcastic:

    Seriously, if you don't intend to overclock, the Phenom is still a decent option.
    If the price is right then why not, but if the price is = then intel wins the market again for a loss in the last Q of 1 billion for amd
Ask a new question

Read More

CPUs Benchmark Phenom Product