- Joined
- Jan 15, 2021
- Messages
- 337 (0.26/day)
So now we compare prices with second hand market. Alrighty...Probably on ebay. With performance dropping below the 6600 why would anyone be interested even if it costs $170...
So now we compare prices with second hand market. Alrighty...Probably on ebay. With performance dropping below the 6600 why would anyone be interested even if it costs $170...
System Name | RogueOne |
---|---|
Processor | Xeon W9-3495x |
Motherboard | ASUS w790E Sage SE |
Cooling | Noctua NH-U14S DX-4677 |
Memory | 128gb Gskill Zeta R5 DDR5 RDIMMs |
Video Card(s) | MSI SUPRIM Liquid X 4090 |
Storage | 1x 2TB WD SN850X | 2x 8TB GAMMIX S70 |
Display(s) | Odyssey OLED G9 (G95SC) |
Case | Thermaltake Core P3 Pro Snow |
Audio Device(s) | Moondrop S8's on schitt Modi+ & Valhalla 2 |
Power Supply | Seasonic Prime TX-1600 |
Mouse | Lamzu Atlantis mini (White) |
Keyboard | Monsgeek M3 Lavender, Akko Crystal Blues |
VR HMD | Quest 3 |
Software | Windows 11 Pro Workstation |
Benchmark Scores | I dont have time for that. |
I mean where are you going to get a 1080ti) new?So now we compare prices with second hand market. Alrighty...
Do you guys know that intel Arc GPUs support ASPM L1 idle power saving mode? Your idle power consumption tests apparently didn't take advantage of this feature.
To TechPowerUp staff:
Do you guys know that intel Arc GPUs support ASPM L1 idle power saving mode? Your idle power consumption tests apparently didn't take advantage of this feature.
It would have been better if they could mention this feature in their articles and give test results from the respective configurations with and without the feature being used.A lot of motherboards don't expose those settings
At least intel's own reference card A770 LE and A750 LE do, but it has not been shown in the comparison chart.Not all of them do
Processor | Ryzen 7 5700X |
---|---|
Memory | 48 GB |
Video Card(s) | RTX 4080 |
Storage | 2x HDD RAID 1, 3x M.2 NVMe |
Display(s) | 30" 2560x1600 + 19" 1280x1024 |
Software | Windows 10 64-bit |
On which univers does the writer of this review finds 1080ti at 170 dollars???
Correct. That's what I do for old SKUs, because they are end-of-life and not produced anymore. They are still relevant for the considerations in this review, at least in my opinion. If you disagree, and only want to buy new, ignore everything relating to these cardsSo now we compare prices with second hand market. Alrighty...
I would assume they do, it's a GPU hardware/driver feature. 99.999% of people have no way to control this or know about it, so it's similar to "why don't you test more undervolting".At least intel's own reference card A770 LE and A750 LE do, but it has not been shown in the comparison chart.
If TPU did test this aspect, it would have helped people know whether this ASRock varient support this feature or not.
Processor | Ryzen 5 7600X |
---|---|
Motherboard | ASRock B650M PG Riptide |
Cooling | Noctua NH-D15 |
Memory | DDR5 6000Mhz CL28 32GB |
Video Card(s) | Nvidia Geforce RTX 3070 Palit GamingPro OC |
Storage | Corsair MP600 Force Series Gen.4 1TB |
3070 is 50% faster, support DLSS and RT, its 4 years newer and cost around $270 so why not? $170 for 7 year old amortized card is terrible idea, just pay 60 more and get brand new 7600, 20% faster with half the power consumptionOn which univers does the writer of this review finds 1080ti at 170 dollars???
People can learn and are curious. You could write about this feature and then the percentage of people who don't know about it or don't use it will decrease.I would assume they do, it's a GPU hardware/driver feature. 99.999% of people have no way to control this or know about it, so it's similar to "why don't you test more undervolting".
Processor | Ryzen 7 5700X |
---|---|
Memory | 48 GB |
Video Card(s) | RTX 4080 |
Storage | 2x HDD RAID 1, 3x M.2 NVMe |
Display(s) | 30" 2560x1600 + 19" 1280x1024 |
Software | Windows 10 64-bit |
it's interesting to see almost every review across the net is testing Arc mainly/only on boards with rBAR on, obviously, because you really need it,To TechPowerUp staff:
Do you guys know that intel Arc GPUs support ASPM L1 idle power saving mode? Your idle power consumption tests apparently didn't take advantage of this feature.
From power testing at the wall, the GPU with ASPM on and displaying ~5-10w, the card is still eating ~40w.on my A770 16GB LE, it looks like this:
View attachment 317190
I have to say though, above ~5W with APSM L1,
vs the usually reported/measured ~40W without,
Doesn't sound logic to me. If this logic is correct, then it would have been 70W idle power consumption when ASPM L1 is not engaged, but the fact is that TechPowerUp has used lab equipment, not by using software readings, to physically measure the total power consumption through the PCI-E power connectors plus the PCI-E bus power lanes, and they actually got the figure of 37 Watts idle.From power testing at the wall, the GPU with ASPM on and displaying ~5-10w, the card is still eating ~40w.
System without A770 using iGPU pulls ~35w at idle.
Add Arc with ASPM working and 5-10w reported in windows and its idling at 70-80w.
The metric the driver is showing is power for the GPU chip only, not the whole card. Meaning that Arc is reporting power like older AMD GPUs. Easiest way to get close to a fully instrumented card for power readings is to add roughly 30W to whatever the software monitor shows you. The other hint that this is the case is the stock power limit, 190w in software on a card officially specced at 225w.
Processor | Ryzen 7 5700X |
---|---|
Memory | 48 GB |
Video Card(s) | RTX 4080 |
Storage | 2x HDD RAID 1, 3x M.2 NVMe |
Display(s) | 30" 2560x1600 + 19" 1280x1024 |
Software | Windows 10 64-bit |
"if the card was cheaper than it is, how much better value would it offer?"On the performance per $ graphs, why are you also showing the 580 at $110, $130 and $150? The narrative makes no mention of these, and I'm not sure what they are intended to show, other than the obvious 'if the card was cheaper than it is then it would offer better value".
"If the card were cheaper than it is, how much better value would it offer?""if the card was cheaper than it is, how much better value would it offer?"
Processor | Ryzen 7 5700X |
---|---|
Memory | 48 GB |
Video Card(s) | RTX 4080 |
Storage | 2x HDD RAID 1, 3x M.2 NVMe |
Display(s) | 30" 2560x1600 + 19" 1280x1024 |
Software | Windows 10 64-bit |
Which may or may not be true. Nobody outside of Intel really knows. and operating at a loss to achieve something else is a reasonable business strategy for a lot of companiesIf Intel supposedly sold this at a loss now
I think that is universally true for microprocessor design and fabricationThey need to enhance [...] design and achieve a better performance/area
We have seen ~70w idle with ASPM off (100-110w at the wall on the 35w baseline system), so its quite likely that in the review power management was working as intended.Doesn't sound logic to me. If this logic is correct, then it would have been 70W idle power consumption when ASPM L1 is not engaged, but the fact is that TechPowerUp has used lab equipment, not by using software readings, to physically measure the total power consumption through the PCI-E power connectors plus the PCI-E bus power lanes, and they actually got the figure of 37 Watts idle.
The 35W (225-190) that should be added to the "GPU chip power draw" readout is under the condition of full load. When idle, VRAM and VRM are not supposed to dissipate that much power.
System Name | Legion |
---|---|
Processor | i7-12700KF |
Motherboard | Asus Z690-Plus TUF Gaming WiFi D5 |
Cooling | Arctic Liquid Freezer 2 240mm AIO |
Memory | PNY MAKO DDR5-6000 C36-36-36-76 |
Video Card(s) | PowerColor Hellhound 6700 XT 12GB |
Storage | WD SN770 512GB m.2, Samsung 980 Pro m.2 2TB |
Display(s) | Acer K272HUL 1440p / 34" MSI MAG341CQ 3440x1440 |
Case | Montech Air X |
Power Supply | Corsair CX750M |
Mouse | Logitech MX Anywhere 25 |
Keyboard | Logitech MX Keys |
Software | Lots |
It's kind of astonishing that they can sell a card with a 406 mm2 die and a 256-bit bus for less than $200 and still make a profit, even if on an older process. NVIDIA and AMD must be swimming in margins.
This is actually a very good card, they have improved the drivers so much. The power draw is horrible, though. Half the performance of a 4070 while drawing more power. It's still a tolerable amount of heat output, but just barely.
It seems that the architecture is just extremely inefficient. The RX 7600 has the same performance as a A770 at half the die size on the same node. That's insane.
What could they achieve with Battlemage? To compete with the upper mid-range they would probably have to get close to 300 W, and for me that's unacceptable no matter how good the price was.
I do hope they stick with it, but I fear they will only be able to compete below my preferred tier of performance.
Another good review. Your attention to detail on this one was excellent! Also agree with the statement that it's a good value at it's current price-point. As far as the cons go, I'm one of those people that prefer the fans be on all the time, even if at minimal speeds. It's a cooling-peace-of-mind thing.Considering the attractive price of just $180, performance is good, too.
When has that never been an option?So now we compare prices with second hand market. Alrighty...
Right?I mean where are you going to get a 1080ti) new?
Considering the card is not sold anymore I realized my comment was super dumb.Another good review. Your attention to detail on this one was excellent! Also agree with the statement that it's a good value at it's current price-point. As far as the cons go, I'm one of those people that prefer the fans be on all the time, even if at minimal speeds. It's a cooling-peace-of-mind thing.
When has that never been an option?
Right?
I wouldn't say "super dumb". It's just that the used market is always an option and part of the buying pool for most people. Comparing only to the new card market ignores a large portion of peoples options.Considering the card is not sold anymore I realized my comment was super dumb.
System Name | ❶ Oooh (2024) ❷ Aaaah (2021) ❸ Ahemm (2017) |
---|---|
Processor | ❶ 5800X3D ❷ i7-9700K ❸ i7-7700K |
Motherboard | ❶ X570-F ❷ Z390-E ❸ Z270-E |
Cooling | ❶ ALFIII 360 ❷ X62 + X72 (GPU mod) ❸ X62 |
Memory | ❶ 32-3600/16 ❷ 32-3200/16 ❸ 16-3200/16 |
Video Card(s) | ❶ 3080 X Trio ❷ 2080TI (AIOmod) ❸ 1080TI |
Storage | ❶ NVME/SSD/HDD ❷ <SAME ❸ SSD/HDD |
Display(s) | ❶ 1440/165/IPS ❷ 1440/144/IPS ❸ 1080/144/IPS |
Case | ❶ BQ Silent 601 ❷ Cors 465X ❸ Frac Mesh C |
Audio Device(s) | ❶ HyperX C2 ❷ HyperX C2 ❸ Logi G432 |
Power Supply | ❶ HX1200 Plat ❷ RM750X ❸ EVGA 650W G2 |
Mouse | ❶ Logi G Pro ❷ Razer Bas V3 ❸ Logi G502 |
Keyboard | ❶ Logi G915 TKL ❷ Anne P2 ❸ Logi G610 |
Software | ❶ Win 10 ❷ 10 ❸ 10 |
Benchmark Scores | I have wrestled bandwidths, Tussled with voltages, Handcuffed Overclocks, Thrown Gigahertz in Jail |
The point of those omissions, I suspect, is to both motivate the user to read other reviews and to save some time for the reviewer.@W1zzard the last 5 GPU reviews (or more) are missing the 7700XT in the performance charts. Can this be amended.
System Name | ❶ Oooh (2024) ❷ Aaaah (2021) ❸ Ahemm (2017) |
---|---|
Processor | ❶ 5800X3D ❷ i7-9700K ❸ i7-7700K |
Motherboard | ❶ X570-F ❷ Z390-E ❸ Z270-E |
Cooling | ❶ ALFIII 360 ❷ X62 + X72 (GPU mod) ❸ X62 |
Memory | ❶ 32-3600/16 ❷ 32-3200/16 ❸ 16-3200/16 |
Video Card(s) | ❶ 3080 X Trio ❷ 2080TI (AIOmod) ❸ 1080TI |
Storage | ❶ NVME/SSD/HDD ❷ <SAME ❸ SSD/HDD |
Display(s) | ❶ 1440/165/IPS ❷ 1440/144/IPS ❸ 1080/144/IPS |
Case | ❶ BQ Silent 601 ❷ Cors 465X ❸ Frac Mesh C |
Audio Device(s) | ❶ HyperX C2 ❷ HyperX C2 ❸ Logi G432 |
Power Supply | ❶ HX1200 Plat ❷ RM750X ❸ EVGA 650W G2 |
Mouse | ❶ Logi G Pro ❷ Razer Bas V3 ❸ Logi G502 |
Keyboard | ❶ Logi G915 TKL ❷ Anne P2 ❸ Logi G610 |
Software | ❶ Win 10 ❷ 10 ❸ 10 |
Benchmark Scores | I have wrestled bandwidths, Tussled with voltages, Handcuffed Overclocks, Thrown Gigahertz in Jail |
The point of those omissions, I suspect, is to both motivate the user to read other reviews and to save some time for the reviewer.