Video summary

Do You Need DDR5 with an RTX 5070? (DDR4 vs DDR5)

Main summary

Key takeaways

Technology

Main question / recommendation

  • The video addresses whether you should buy DDR5 or DDR4 when pairing with an RTX 5070, framing the decision as a “rip-off” due to current RAM pricing.
  • Recommendation: Choose DDR4 right now, because DDR5 is far more expensive than DDR4—both in absolute cost and as a relative price increase.
  • A key workaround is highlighted: certain Intel platforms (the video names 12th, 13th, and 14th-gen Intel CPUs) can support either DDR4 or DDR5 depending on the motherboard.

Platform / CPU value points

  • The speaker favors Intel 12th gen, specifically the i7-12700, as strong value.
  • Claimed value: i7 models are often under $300, and have been seen well under $200 (example: i7 12700).

Benchmark setup (what was tested)

GPU

  • RTX 5070
  • Used with undervolting for lower power bills.
  • The video notes this is not intended as an undervolting tutorial.

Resolutions and settings

  • 1080p: Low settings
  • 1440p: Medium settings
  • 4K: High settings

RAM kits used

  • DDR4: 3600 CL18
  • DDR5: 6000 CL34

Games tested (6)

  1. Counter-Strike 2
  2. Resident Evil (subtitled as “Resident Evil 9”)
  3. Pragmata
  4. Marvel Rivals
  5. Fortnite
  6. 007 First Light

Results summary: DDR4 vs DDR5

Overall theme: Differences are usually small, and often not worth a large DDR5 price premium. DDR5 sometimes wins—especially in 0.1% lows—but the size of the gains generally isn’t dramatic enough to justify the extra cost.

Counter-Strike 2

  • 1080p (low): virtually no difference
  • 1440p: slight advantage for DDR5 (small improvement)
  • 4K: DDR4 shows ~15% difference (noted as surprisingly large)

Takeaway: At competitive/low settings and 1440p, gaps are minimal.

Resident Evil

  • 1080p: DDR5 shows a large advantage (example numbers: approx 166 avg vs 144 avg FPS)
  • 1440p: gap shrinks a lot (121 vs 118)
  • 4K: nearly identical (about 1 FPS difference)

Takeaway: Larger gap at 1080p, but performance converges at higher resolutions.

Pragmata

  • 1080p: about 3–4% difference (DDR5 vs DDR4)
  • 1440p: differences reduce to about ~2 FPS
  • 4K: no average FPS difference

Takeaway: RAM impact scales down as resolution increases.

Marvel Rivals

  • 1080p: some difference in average FPS
  • 1440p: virtually no difference
  • 4K: differences mostly show up in 0.1% lows, while averages are near identical

Fortnite (best example of “no difference”)

  • 1080p: small average difference (~226 vs 221 FPS)
  • 1440p: no meaningful difference
  • 4K: minimal difference

Note: The speaker argues Fortnite is highly optimized, which explains its low RAM sensitivity.

007 First Light

  • 1080p: slight DDR5 advantage (~137 vs 142 FPS)
  • 0.1% lows: generally better on DDR5, but not “game-breaking”
  • 1440p: small average difference (~2 FPS)
  • 4K: averages match, but 0.1% lows vary slightly (DDR4 vs DDR5 depending on the segment)

Methodological note: Because 0.1% lows are highly sensitive, the speaker says results may vary. They ran three benchmarks and averaged the 0.1% lows.

0.1% lows (stability) and power/CPU behavior

  • The speaker repeatedly notes: 0.1% low FPS often favors DDR5, suggesting smoother worst-case performance at times.
  • However, they argue the overall experience isn’t “ruined” by DDR4 when averages are close.
  • Additional benefit attributed to DDR4: the CPU uses less power with DDR4 than DDR5 across the benchmarks.
    • This matters if you’re undervolting the RTX card—the video claims DDR4 can further reduce system power.

Cost/price analysis (why they say DDR4 is the better buy)

  • The video concludes DDR5’s current ~$300 price differential isn’t justified by measured performance gains.
  • Comparison to “normal” pricing:
    • Over a year ago, the DDR5 vs DDR4 differential might have been around ~$40.
    • At that time, performance gains (sometimes several percent or more) were easier to justify.
  • Current situation:
    • On a $1,000 build, adding $300 is described as roughly a ~30% cost increase (considering RAM/platform considerations), making DDR5 harder to recommend.
  • Another value pressure point mentioned: the RTX 5070 is more expensive than before (example claim: $600 up to ~$850+), further squeezing the “value” argument.

Testing philosophy / future plans

  • The speaker says they used to test higher-end GPUs like RTX 5090, but plan to stop unless requested because:
    • The price is “absurd”
    • Power consumption isn’t realistic for typical users
  • They personally preferred 5080 over 5090 earlier due to practicality and power draw.

Main speakers / sources

  • Primary speaker: The video narrator/reviewer (unnamed in the subtitles), running DDR4 vs DDR5 gaming benchmarks with RTX 5070 on a flexible Intel platform that supports DDR4 or DDR5 using 12th/13th/14th-gen CPUs.
  • Sponsor mentioned: VIP SCD Keys (Windows 11 license keys).

Original video