HardwareCanucks TitanZ > R9 295X2

[Arkadrel]

Logically it should be slower, yeah the review must be biased.

Review payed for by nvidia 😛 jk


Not sure what to say to HardwareCanucks, it looks bad though.

 

[KaRLiToS]

Yeah, this is looking screwy. I normally like these guys, but either something is horribly wrong, they got a golden card, there are new NVidia drivers, or the card was cherry-picked.

Or maybe they just falsified the data. It is really easy to make a graph.

Even if it was a cherry picked card, the card is still a dual GPUs card and only has a single fan.

Titan Z is a beast!

Just put it underwater and it will be the fastest 4K kid out there...

Under water it must be very fast.

You can always buy this nice sku. EVGA 12G-P4-3999-KR GeForce GTX TITAN Z 12GB 768-Bit GDDR5 PCI Express 3.0 SLI Support Hydro Copper Video Card


Only 3,399$ with a waterblock.

Usually, a dual GPU block is around 180$-200$ , see here the price for the R9 295x2 blocks

But when it is a nVidia dual GPUs, it's a 399$ block. 🙄:thumbsdown::thumbsdown:

This is ridiculous.

The card must be awsome especially with a waterblock but honestly, they are really screwing everything up with the prices.

 

[Tempered81]

Hardware Canucks = Nvidia shill, just like PCPer... fanboys. Hardware Canucks has historically produced the most biased review percentages - and not much has changed. I wouldn't expect anything less from then when it comes to Titan Z vs. 295X2

Take a look at GTX480 launch reviews - Notice how skewed Hardware Canucks' percentages are compared to all other major review sites... even though these reviews all came out within 1 or 2 days of each other. Canucks gives them a default 10% performance lead.

 

[Flapdrol1337]

lol, falsified data, that's ridiculous.

In other reviews the 295x2 is what, on average 7% faster on 1440p? The hardwarecanucks card boosts 100 MHz higher than average, which explains the results completely. It's just a good card or good ventilation.

 

[Silverforce11]

Actually, 960-980mhz is much higher than ~700mhz that other sites have it at once its warmed up.

Work out the % difference, its huge.

We all knew that single fan wasn't capable to cool twin GK110 with that much power consumption and other sites were spot on.. all except HWC, with silent fan speeds and much higher boost clocks.

How? Did the reviewer crank up fan speed for benches while testing sound at default? Toms actually did this on their 780 review but they stated it was done on purpose (80% manual fan during benches) to prevent the 780 from throttling below its top boost bin. HWC didn't specify.

 

[ShintaiDK]

Remember the 295X2 cant hold its boost clock either. Even with its dual air and water cooling, AC in room and without the constrains of a chassis.

http://www.techpowerup.com/reviews/AMD/R9_295_X2/29.html

Seems its all about quickly testing cards today from both vendors, so they can cool down again before throttle 😉

 

[Leadbox]

Remember the 295X2 cant hold its boost clock either. Even with its dual air and water cooling, AC in room and without the constrains of a chassis.

http://www.techpowerup.com/reviews/AMD/R9_295_X2/29.html

Seems its all about quickly testing cards today from both vendors, so they can cool down again before throttle 😉

Was that a game tested or were they using their "mystery app" :ninja:

 

[3DVagabond]

Remember the 295X2 cant hold its boost clock either. Even with its dual air and water cooling, AC in room and without the constrains of a chassis.

http://www.techpowerup.com/reviews/AMD/R9_295_X2/29.html

Seems its all about quickly testing cards today from both vendors, so they can cool down again before throttle 😉

Why so secretive about what he uses to load the card? When asked directly, this is what he said.

not furmark, but something that mimics the peak load of the most demanding games

So it was with a synthetic load. Might explain why nobody else suffered throttling while testing with games.

 

[ShintaiDK]

Why so secretive about what he uses to load the card? When asked directly, this is what he said.



So it was with a synthetic load. Might explain why nobody else suffered throttling while testing with games.

I heard all those excuses before. If your test is short enough they dont throttle. Or using something like 18C ambient that nVidia and AMD prefers to be tested in. The 290/290X for example was first "fixed" with custom designs. Even the regular Titan throttles, tho less than the 780ti.

Its just the new black from both companies. To demo their cards in situations the enduser as such cant use for anything. But again, with shinking sales and slow progress, deperation starts to kick in.

 

[3DVagabond]

I heard all those excuses before. If your test is short enough they dont throttle. Or using something like 18C ambient that nVidia and AMD prefers to be tested in. The 290/290X for example was first "fixed" with custom designs.

Its just the new black from both companies. To demo their cards in situations the enduser as such cant use for anything. But again, with shinking sales and slow progress, deperation starts to kick in.

Which is why I prefer [H]'s methods.

 

[Techhog]

lol, falsified data, that's ridiculous.

In other reviews the 295x2 is what, on average 7% faster on 1440p? The hardwarecanucks card boosts 100 MHz higher than average, which explains the results completely. It's just a good card or good ventilation.

The thing that needs to be explained is how the difference is so enormous. It's really weird. Most sites use an open testbench, so ventilation is out.

 

[bakalu]

PCPER and Guru3D used 14.6 drivers for the R9 295X2 and the Hardware Canucks review used 14.4 drivers.

Hardware Canucks review also used 14.6 drivers.

Drivers:
AMD 14.6 Beta
NVIDIA 337.88 Beta

*Please note that a previous version of this page erroneously listed AMD's 14.4 driver. All testing on AMD hardware was conducted with the 14.6 beta driver.

http://www.hardwarecanucks.com/foru...6869-nvidia-titan-z-performance-review-4.html

 

[blackened23]

We're still working on our comprehensive list of shill sites, right? So hardwarecanucks paid by NV, while hardOCP is paid fully AMD (but were paid by nvidia during the 600 series), and then you have "AMD zone/center" anandtech, PCPer who goes back and forth depending on the day of the week - I think they're paid by nvidia monday through wednesday but AMD pays them for reviews on Thursdays and Fridays. I think TPU is paid every on even months by NV, and odd months by AMD.

Still working on the comprehensive list here.

On a related note, did anyone read the conclusion of the hardwarecanucks review? Seems that their recommendation was to get the 295X2 instead of the Titan Z. That's a hell of a conclusion to make in a review that was paid for by nvidia! 😱 LOL

 

[Techhog]

We're still working on our comprehensive list of shill sites, right? So hardwarecanucks paid by NV, while hardOCP is paid fully AMD (but were paid by nvidia during the 600 series), and then you have "AMD zone/center" anandtech, PCPer who goes back and forth depending on the day of the week - I think they're paid by nvidia monday through wednesday but AMD pays them for reviews on Thursdays and Fridays. I think TPU is paid every on even months by NV, and odd months by AMD.

Still working on the comprehensive list here.

On a related note, did anyone read the conclusion of the hardwarecanucks review? Seems that their recommendation was to get the 295X2 instead of the Titan Z. That's a hell of a conclusion to make in a review that was paid for by nvidia! 😱 LOL

Yeah, I don't think that they're shills. However, I have to really doubt their testing methods. Either way, I expect this will become the go-to review for Nvidia fanboys.

 

[SlowSpyder]

I heard all those excuses before. If your test is short enough they dont throttle. Or using something like 18C ambient that nVidia and AMD prefers to be tested in. The 290/290X for example was first "fixed" with custom designs. Even the regular Titan throttles, tho less than the 780ti.

Its just the new black from both companies. To demo their cards in situations the enduser as such cant use for anything. But again, with shinking sales and slow progress, deperation starts to kick in.

Other reviews from the the R9 295 thread don't mention such issues.

Anandtech says the R9 295 barely throttles.

With the exception of our strategy games, both of which throttle at certain points due to power restrictions on both the R9 295X2 and R9 290X, AMD’s latest dual-GPU card is able to maintain 1018MHz on all of our games.

Forbes.

With the 295X2′s implementation of hybrid liquid and air cooling, that’s a thing of the past. This time around AMD promised “up to” a 1018MHz GPU clock, and AMD delivered a constant 1018MHz GPU clock with no dips in performance. Whether it’s a few minutes or a few hours, you’re getting 1018MHz. I tested this using a 2 hour Furmark stress test as well as an overnight session of Unigine’s Heaven. That clock speed was unwavering.

HardOCP

In our testing we found that the cooling mechanisms all work well and keep the GPU temps under 70c. More importantly, it allows the GPUs to keep a consistent 1018MHz clock speed while gaming.

Tom's

I didn't see throttling specifically mentioned one way or the other, but they do mention the temps never went above 65C, so I doubt it throttled to any significant degree.

The Asetek closed-loop liquid cooler does a fairly good job, though it's only equipped with a 120 mm radiator. A peak GPU temperature of 65 °C is admirable, particularly considering that we couldn't push a single Radeon R9 290X below 50 °C in our aftermarket cooling project.

I'd be surprised if throttling was an issue for the R9 295.

 

[Techhog]

Other reviews from the the R9 295 thread don't mention such issues.

Anandtech says the R9 295 barely throttles.




Forbes.




HardOCP




Tom's

I didn't see throttling specifically mentioned one way or the other, but they do mention the temps never went above 65C, so I doubt it throttled to any significant degree.




I'd be surprised if throttling was an issue for the R9 295.

I've heard that there's power limit throttling (they really should have added a third 8-pin), but that's all. That wouldn't affect most games if you have a fantastic PSU. The testing environment and length would have little to no effect.

 

[caswow]

another funny thing. titan z has only 2 x 8 pins. amd got shit on for their choice but no one had mentioned this yet.

 

[KaRLiToS]

So according to SKYMTL, he updated the review cause he actually tested with 14.6 beta.

I have updated the review actually. It was a typo on that page.

And in the review:

Drivers:
AMD 14.6 Beta
NVIDIA 337.88 Beta

*Please note that a previous version of this page erroneously listed AMD's 14.4 driver. All testing on AMD hardware was conducted with the 14.6 beta driver.

 

[ShintaiDK]

Other reviews from the the R9 295 thread don't mention such issues.

Anandtech says the R9 295 barely throttles.




Forbes.




HardOCP




Tom's

I didn't see throttling specifically mentioned one way or the other, but they do mention the temps never went above 65C, so I doubt it throttled to any significant degree.




I'd be surprised if throttling was an issue for the R9 295.

Its funny that for example that forbes "review" and anandtech completely contradicts one another. But again, maybe easy to miss when you want a predefined outcome. Lets just be honest about it, both companies clock their cards higher than they in reality can handle in prolonged gaming to get better benchmark results.

 

[Gikaseixas]

If it looks like a duck, quacks like a duck and walks like a duck, it's a duck"

People can try as hard as they want to spin this one but they'll fail since the 295X is simply the better, faster card period.

 

[Gikaseixas]

On a related note, did anyone read the conclusion of the hardwarecanucks review? Seems that their recommendation was to get the 295X2 instead of the Titan Z. That's a hell of a conclusion to make in a review that was paid for by nvidia! LOL

The recommendation alone is not enough. People who actually read the review will get a clouded picture due to the abnormal numbers they gave to the Titan Z.
That's what people are talking about here

 

[SlowSpyder]

Its funny that for example that forbes "review" and anandtech completely contradicts one another. But again, maybe easy to miss when you want a predefined outcome. Lets just be honest about it, both companies clock their cards higher than they in reality can handle in prolonged gaming to get better benchmark results.

Want a predefined outcome? Who? Me? Forbes? Anandtech?

I don't own either card or even a single GPU card based on Hawaii or big Kepler, so I can't say much about products that I don't own. Not all reviews agree with one another on every single detail. The review this thread is about is an example of that. The Forbes review says that the card never throttled. The Anandtech review said it only throttled a little. Hardocp said they warm there cards up for a minimum of 15 minutes before they bench, and again no throttling.

I couldn't tell you much beyond that, and certainly nothing on personal experience, but most review sites seem to suggest little to no throttling. But of course every case and it's airflow and environment will be different.

 

[ShintaiDK]

I couldn't tell you much beyond that, and certainly nothing on personal experience, but most review sites seem to suggest little to no throttling. But of course every case and it's airflow and environment will be different.

And golden samples, special press cards with special BIOS etc.

Its just a area with little trust left.

 

[Gikaseixas]

And golden samples, special press cards with special BIOS etc.

Its just a area with little trust left.

HWC sample is special then, it outperforms GTX 780ti SLI :sneaky:
And yet people defend this D:

 

[96Firebird]

HWC sample is special then, it outperforms GTX 780ti SLI :sneaky:
And yet people defend this D:

I just glanced at the review, conclusion says 780 Ti SLI is 6% faster at 4K, same performance at 1440p. Quickly looked at the charts, when TitanZ is faster, it is within 5% of 780 Ti SLI. Surely you know there is a variance in the result, just look at the graphs and it will give you an idea.

Unless I am missing something...?