Serious Sam 2 continues on with the idea that
the more enemies on screen at once the better. This also
brings complexity to the game as there are more things to render
both graphically and physically. The game is VS3 and PS3
compatible so it can take what current video cards are able to
handle. Lets see how the VS3 based x1650 compares to the
VS2 x800XL in this test.
|
1600 AA - Results (In FPS)
|
Minimum |
Average |
Maximum |
| Asus Extreme AX800XL: |
20 |
26.42 |
44 |
| HIS x1650 Pro: |
16 |
21.52 |
34 |
| HIS x1650 Pro 658/760MHz:: |
17 |
22.97 |
35 |
This resolution and quality
setting brings both cards to their knees, as neither one is
what I'd consider playable. Again the x800XL gives an
extra 22% improvement over the stock speeds of the x1650, while
overclocking the x1650 gives you almost a 7% increase (1.5fps)
over stock speeds. Again the graphs here a very flat but
the average frame rate doesn't allow a playable gaming experience
in this case. Lets see what other games do with this newer
card.
Game Tests - F.E.A.R.
Next is F.E.A.R, a fps that
offers a high quality gaming experience, with the added enjoyment
of 'bullet time'. The game itself isn't that interesting
to me, but the engine offers some higher graphical quality so
lets see what it does to our test cards.

|
1024 NAA - Results (In FPS)
|
Minimum |
Average |
Maximum |
| Asus Extreme AX800XL: |
26 |
59.82 |
138 |
| HIS x1650 Pro: |
19 |
50.18 |
109 |
We can see similar
results to that of Serious Sam 2, with the higher fillrate/bandwidth
based x800XL performing noticeably faster (19%) than the newer
card. That isn't to say that both cards aren't playable
at this setting, as they are but the x1650 is somewhat slower
than its older but higher end brother. Will adding AA
and anisotropic filtering change anything?

|
1024 AA - Results (In FPS)
|
Minimum |
Average |
Maximum |
| Asus Extreme AX800XL: |
20 |
38.93 |
96 |
| HIS x1650 Pro: |
16 |
38.55 |
94 |
| HIS x1650 Pro 658/760MHz:: |
15 |
41.29 |
99 |
Here things are much closer
than the previous results, with the two cards being within
1% of each other, or otherwise a tie. Overclocking the
card gives a 7% increase which has been pretty consistent
across our test results for the x1650. Again all cards
are playable, though the x1650 does dip below 20fps but only
for two seconds at the start of the test. Lets pump
up the resolution.

|
1600 NAA - Results (In FPS)
|
Minimum |
Average |
Maximum |
| Asus Extreme AX800XL: |
17 |
30.86 |
62 |
| HIS x1650 Pro: |
12 |
26.34 |
58 |
The results match those of
the non-AA results in this game with the x800XL being 17%
faster, though in this case it equates to only 4fps on average.
The x1650 isn't on my list of playable cards at this resolution,
as it spends most of its time in the very low 20fps area.
The x800XL is right on the cusp of playability but I'd probably
prefer not to play at this resolution with these cards.

|
1600 AA - Results (In FPS)
|
Minimum |
Average |
Maximum |
| Asus Extreme AX800XL: |
11 |
19.21 |
45 |
| HIS x1650 Pro: |
9 |
17.88 |
39 |
| HIS x1650 Pro 658/760MHz:: |
10 |
18.73 |
40 |
Unlike the 1024 AA results
for FEAR, the x1650 doesn't quite catch up to the x800XL.
This is a rather moot point as both cards are very much in
the unplayable category for this test. As a note, overclocking
the x1650 only gives a ~5% boost in performance, slightly
lower than the 7% we have seen before.
Game Tests - Lost Coast
Lastly we will look at Half
Life 2: Lost Coast, which was a the forefront of the HDR experience,
in fact this level is really an HDR tester. Most of the
features from this made it directly to Episode 1, so performance
should be very similar between these two games. Lets see
the results.

|
1024 NAA - Results (In FPS)
|
Minimum |
Average |
Maximum |
| Asus Extreme AX800XL: |
34 |
58.12 |
106 |
| HIS x1650 Pro: |
36 |
64.79 |
116 |
This is the first game since
Quake 4 where the x1650 has a lead, let alone a large one.
There is a 11% performance increase by moving to this card.
The difference is noticeable in the graph as the speed up
starts at the 22 second mark and goes from there. Both
cards are very playable at this resolution but what about
adding to the quality settings?

|
1024 AA - Results (In FPS)
|
Minimum |
Average |
Maximum |
| Asus Extreme AX800XL: |
20 |
37.81 |
63 |
| HIS x1650 Pro: |
29 |
49.55 |
96 |
| HIS x1650 Pro 658/760MHz:: |
31 |
51.20 |
101 |
At these settings there is
a very marked difference between the two cards, on the order
of 31%. This seems quite suspious but the tests were
very repeatable, though if I notice any difference with the
Cat 6.9 drivers, I will update this review. Overclocking
the x1650 only gives a 3.3% increase in frame rate, though
it doesn't change that the x1650 is playable at this setting.
The x800XL is pretty playable here and I would find it just
fine to play the game. Lets bump up the resolution for
the last time.

|
1600 NAA - Results (In FPS)
|
Minimum |
Average |
Maximum |
| Asus Extreme AX800XL: |
19 |
33.47 |
50 |
| HIS x1650 Pro: |
25 |
46.03 |
88 |
Again we can see quite a
difference between the two cards, this time on the order of
37% faster performance. The x1650 is very playable again,
while the x800XL is probably playable, but is very close to
the 30fps mark which I find a marker for myself as to the
playability of a game. Lets see if the domination by
the x1650 in this test continues in our last test setting.

|
1024 AA - Results (In FPS)
|
Minimum |
Average |
Maximum |
| Asus Extreme AX800XL: |
10 |
18.31 |
32 |
| HIS x1650 Pro: |
17 |
31.05 |
62 |
| HIS x1650 Pro 658/760MHz:: |
18 |
32.60 |
67 |
We again see the large difference
between the two cards in our Half Life 2 test. The x1650
is almost 70% faster than the x800XL here and in this case
it makes a difference, with the x1650 moving into the low
end of what I'd consider playable, not my recommend average
frame rate but it can do. Overclocking gains a 5% increase
in frame rate, much like we saw earlier in the review.
One last test we ran has to
do with ATi's AVIVO video, a feature found only only on the
x1K series of video cards. So I ran a few HD videos, some
720p and one 1080p video on my computer to see the respective
CPU loads. As a note the CPU had Hyperthreading enabled,
and the desktop resolution was 1600*1200 for all but the 1080p
video where the resolution was 1800*1440 the maximum of my monitor.
We used a 720p x264 video, a 720p WMV (Flight), and the same
WMV video but in 1080p. So lets see if AVIVO does anything
to lessen the CPU load.
| |
Asus AX800XL-2DTV |
HIS x1650 Pro IceQ Turbo
Dual DL-DVI |
| x264 720p Video: |
15% |
13% |
| WMV 720p Video: |
28% |
25% |
| WMV 1080p Video: |
50% |
33% |
These are not exact numbers,
as they are taken during the playback of the video and are
just eyeballed. However the 1080p video shows some interesting
data, at this setting a single core PIV 2.8GHz would be completely
loaded down with playing the video on average, so there would
be some dropped frames without any assistance from the video
card. The AVIVO based card lowers the CPU usage to something
much easier to handle by the CPU without dropped frames.
Conclusion
HIS
has provided us with one of the newest cards on the market,
due to their being a ATi partner, but did this card have what
it takes to be a market leader in its segment?
First the box and its contents.
The box itself catches your eye and the plastic window allows
you to see the card, and its cooling fan without having to take
it out of the box. The contents are a good mix, the software
is better than some in the video/image editing area, though
without a video-in this is a slightly odd inclusion. Games
were mainly missing, but most bundles aren't anything to write
home about, as long as everything else gains from the money
saved there.
The video card itself is dominated
by the IceQ cooler, which although big isn't at all heavy in
comparison to some other video cards. Cooling was well
taken care of with that cooler as well as the ramsinks.
This allowed the card to keep cooler than many other cards I've
looked at, while still keeping very quiet in overclocked operation.
Overclocking wasn't anything to write home about but there was
a increase that we were able to find out of this card.
2D quality is something that
ATi and its partners can be very proud of, as it is among the
best I've ever seen. Game results were a mixed bag, though
for the most part it kept up with the x800XL an older but higher
end card. The 1600*1200 results in Serious Sam 2 and the
1024*768 NAA F.E.A.R results were the exception where the x800XL
beat the x1650 by a wide (~20%) margin. However if you
play Half Life 2 or any of its newer variants I would definitely
recommend the x1650 over the x800XL as it beats it from 11%-70%.
Overall unless you play Half Life 2 a fair amount you would
find it and the x800XL good for playing at most of the same
resolutions.
Price is always a factor in
buying a video card, at this time I couldn't find any direct
pricing for this card, but it should be around the price of
the older x1600XT which is basically the same unit which goes
for about $130(US). Considering the price of the x800XL
is still in the $250(US) range, if you can find this card for
around the same price as the x1600XT and you don't need 1600*1200
gaming, but want a very solid Media Center based card, then
this is a very good idea for you. If the price is too
much above that $150(US) mark I would go with a x1600XT with
similar features which should be cheaper. As a bonus,
if you have a CrossFire capable board you can get a second one
later for increased performance in most games. I would
like to thank HIS
for providing this card for us to review.